In [2]:
# Python ≥3.5 is required
import sys
assert sys.version_info >= (3, 5)
# Scikit-Learn ≥0.20 is required
import sklearn
assert sklearn.__version__ >= "0.20"
# TensorFlow ≥2.0 is required
import tensorflow as tf
from tensorflow import keras
assert tf.__version__ >= "2.0"
# Common imports
import numpy as np
import os
# To plot pretty figures
%matplotlib inline
import matplotlib as mpl
import matplotlib.pyplot as plt
mpl.rc('axes', labelsize=14)
mpl.rc('xtick', labelsize=12)
mpl.rc('ytick', labelsize=12)
import pandas as pd
Load data¶
In [3]:
def readtimes(file):
times = []; values = []
with open(file) as f:
for line in f:
if 'groups' in line:
groups = line.replace(',',' ').split()[5:]
if 'TIME' in line:
newtime = True
break
while newtime:
newtime = False
times.append(float(line.replace(',',' ').split()[2]))
values_per_time = []
for line in f:
if 'GROUP' in line:
continue
if 'TIME' in line:
newtime = True
break
values_per_time.append([float(x) for x in line.strip().split(',') if x != ''])
values.append(values_per_time)
return times, groups, values
In [4]:
times, groups, values = readtimes('group-output-time_mod.csv')
In [5]:
times = np.array(times)
values = np.array(values)
print('# of times: ', len(times))
print('# of groups: ', len(groups))
print('values.shape: ', values.shape)
# of times: 3888 # of groups: 8 values.shape: (3888, 80, 10)
In [6]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(values[3090,i*10:(i+1)*10,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [7]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, values[:,i*10,4])
ax.flatten()[i].set_title(group)
Prepare the data¶
In [8]:
nl = int(values.shape[-1])
nc = int(values.shape[-2]/len(groups))
print('Grid: ', nl, 'x',nc)
Grid: 10 x 10
In [9]:
X_train_2D = values
In [10]:
X_train_2D.shape
Out[10]:
(3888, 80, 10)
In [11]:
X_train_3D = X_train_2D.reshape(len(times),len(groups),nl,nc)
In [12]:
X_train_3D.shape
Out[12]:
(3888, 8, 10, 10)
In [13]:
X_train_1D = X_train_2D.reshape(len(times),len(groups)*nl*nc)
In [14]:
X_train_1D.shape
Out[14]:
(3888, 800)
In [15]:
from sklearn.preprocessing import StandardScaler
stdscaler = StandardScaler()
X_train_1D_norm = stdscaler.fit_transform(X_train_1D)
X_train_2D_norm = X_train_1D_norm.reshape(len(times),len(groups)*nl, nc)
X_train_3D_norm = X_train_1D_norm.reshape(len(times),len(groups),nl,nc)
In [16]:
def calculateerror(X_train_3D, X_train_3D_recovered, groups, print_step = 0):
abs_error = abs(X_train_3D - X_train_3D_recovered)
perc_error = abs_error*100/abs(X_train_3D)
print('max_abs_error: ',np.max(abs(X_train_3D - X_train_3D_recovered)) )
print('mean_abs_error: ',np.mean(abs(X_train_3D - X_train_3D_recovered)) )
if print_step:
for time in range(0,X_train_3D.shape[0],print_step):
print('\ntime: ',time)
for i, group in enumerate(groups):
print('Group '+group+': max_abs_error = ',
round(np.max(abs_error[time,i,:,:]) ,4),
' %_mae = ',
round( np.max(perc_error[time,i,::][np.isfinite(perc_error[time,i,::])]) ,4),
'%')
Dimensionality reduction - PCA¶
In [17]:
from sklearn.decomposition import PCA
pca = PCA(X_train_1D.shape[1])
X_train_pca = pca.fit_transform(X_train_1D)
In [18]:
X_recovered = pca.inverse_transform(X_train_pca)
np.allclose(X_recovered, X_train_1D)
Out[18]:
True
In [19]:
#print(pca.singular_values_**2/(X_train_1D.shape[0]-1))
#print()
#print(pca.explained_variance_)
#print(pca.explained_variance_ratio_)
In [20]:
p = 0.999
cumsum_eig = np.cumsum(pca.explained_variance_ratio_)
d = np.argmax(cumsum_eig >= p) + 1
d
Out[20]:
6
In [21]:
plt.figure(figsize=(6,4))
plt.plot(cumsum_eig, linewidth=3)
plt.xlabel("Dimensions")
plt.ylabel("Explained Variance")
plt.ylim([cumsum_eig[0],1.1])
plt.plot([d, d], [0, p], "k:")
plt.plot([0, d], [p, p], "k:")
plt.plot(d, p, "ko")
plt.annotate("Elbow", xy=(d, p), xytext=(d, cumsum_eig[0]+0.05),
arrowprops=dict(arrowstyle="->"), fontsize=16)
plt.grid(True)
plt.show()
In [22]:
p = 0.999
cumsum_sv = np.cumsum(pca.singular_values_/sum(pca.singular_values_))
d = np.argmax(cumsum_sv >= p) + 1
d
Out[22]:
15
In [23]:
plt.figure(figsize=(6,4))
plt.plot(cumsum_sv, linewidth=3)
plt.xlabel("Dimensions")
plt.ylabel("Cumsum Singular values")
plt.ylim([cumsum_sv[0],1.1])
plt.plot([d, d], [0, p], "k:")
plt.plot([0, d], [p, p], "k:")
plt.plot(d, p, "ko")
plt.annotate("Elbow", xy=(d, p), xytext=(d, cumsum_sv[0]+0.05),
arrowprops=dict(arrowstyle="->"), fontsize=16)
plt.grid(True)
plt.savefig('pca_normCumSum_singularValues.png')
plt.show()
Build the model - PCA¶
In [24]:
pca_compress = PCA(n_components=15)
X_train_pca = pca_compress.fit_transform(X_train_1D)
X_recovered = pca_compress.inverse_transform(X_train_pca)
np.allclose(X_recovered, X_train_1D)
Out[24]:
False
In [25]:
fig, ax = plt.subplots(1,1, figsize=[20,10])
ax.plot(times, X_train_pca);
ax.grid()
ax.legend(range(15))
Out[25]:
<matplotlib.legend.Legend at 0x7faff5a0d710>
In [26]:
import joblib
joblib.dump(pca_compress, "pca_compress_15.pkl")
np.savetxt('X_train_1D.csv', X_train_1D, delimiter=',')
np.savetxt('X_train_pca.csv', X_train_pca, delimiter=',')
np.savetxt('times.csv', times, delimiter=',')
with open('groups.txt','w') as f:
f.writelines([g + '\n' for g in groups])
#...
# pca_compress = joblib.load("pca_compress_15.pkl")
# X_train_compressed = np.loadtxt('X_train_pca.csv', delimiter=',')
# X_train_1D = np.loadtxt('X_train_1D.csv', delimiter=',')
# times = np.loadtxt('times.csv', delimiter=',')
# with open('groups.txt') as f:
# groups = [g.strip() for g in f.readlines()]
# # PCA recovered
# X_recovered = pca_compress.inverse_transform(X_train_compressed)
In [27]:
calculateerror(X_train_1D.reshape(len(times),len(groups),nl,nc),
X_recovered.reshape(len(times),len(groups),nl,nc),
groups,
print_step=0)
max_abs_error: 0.9455268950468039 mean_abs_error: 0.003775904514174
/home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: divide by zero encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until /home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until
In [28]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(X_train_1D.reshape(len(times),len(groups),nl,nc)[100,i,:,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [29]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(X_recovered.reshape(len(times),len(groups),nl,nc)[100,i,:,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [30]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_train_1D[:,i*nl*nc+4])
ax.flatten()[i].set_title(group)
In [31]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_recovered[:,i*nl*nc+4])
ax.flatten()[i].set_title(group)
In [32]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_train_1D[:,i*nl*nc+4])
ax.flatten()[i].plot(times, X_recovered[:,i*nl*nc+4],'--')
ax.flatten()[i].set_title(group)
plt.savefig('pca_compression.png')
Dimensionality reduction - DEEP AUTOENCODER¶
Linear autoencoder¶
In [30]:
np.random.seed(42)
tf.random.set_seed(42)
# Need to have validation loss
early_stopping = keras.callbacks.EarlyStopping(monitor='val_loss',
min_delta=0.0,
patience=100,
verbose=2,
restore_best_weights=True)
encoder = keras.models.Sequential([keras.layers.Dense(15, input_shape=[800])])
decoder = keras.models.Sequential([keras.layers.Dense(800, input_shape=[15])])
autoencoder = keras.models.Sequential([encoder, decoder])
autoencoder.compile(loss="mse",
optimizer=keras.optimizers.Nadam(lr=0.0007, beta_1=0.9, beta_2=0.999)
)
autoencoder.summary()
Model: "sequential_2" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= sequential (Sequential) (None, 15) 12015 _________________________________________________________________ sequential_1 (Sequential) (None, 800) 12800 ================================================================= Total params: 24,815 Trainable params: 24,815 Non-trainable params: 0 _________________________________________________________________
In [31]:
history = autoencoder.fit(X_train_1D_norm,
X_train_1D_norm,
epochs=1000,
validation_data=(X_train_1D_norm, X_train_1D_norm),
callbacks=[early_stopping])
Train on 3888 samples, validate on 3888 samples Epoch 1/1000 3888/3888 [==============================] - 1s 284us/sample - loss: 0.0789 - val_loss: 0.0311 Epoch 2/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 0.0196 - val_loss: 0.0143 Epoch 3/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 0.0112 - val_loss: 0.0080 Epoch 4/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.0058 - val_loss: 0.0043 Epoch 5/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.0034 - val_loss: 0.0028 Epoch 6/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 0.0021 - val_loss: 0.0019 Epoch 7/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 0.0019 - val_loss: 0.0018 Epoch 8/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 0.0018 - val_loss: 0.0017 Epoch 9/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 0.0017 - val_loss: 0.0017 Epoch 10/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 0.0016 - val_loss: 0.0017 Epoch 11/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 0.0015 - val_loss: 0.0014 Epoch 12/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 0.0013 - val_loss: 0.0013 Epoch 13/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.0012 - val_loss: 0.0010 Epoch 14/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 9.3700e-04 - val_loss: 8.1995e-04 Epoch 15/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.5246e-04 - val_loss: 6.6059e-04 Epoch 16/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.5002e-04 - val_loss: 5.4606e-04 Epoch 17/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.1014e-04 - val_loss: 4.6631e-04 Epoch 18/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.3811e-04 - val_loss: 4.0917e-04 Epoch 19/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 4.6926e-04 - val_loss: 3.6424e-04 Epoch 20/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 3.3174e-04 - val_loss: 2.9551e-04 Epoch 21/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 2.5156e-04 - val_loss: 2.0485e-04 Epoch 22/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.7204e-04 - val_loss: 1.3412e-04 Epoch 23/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 1.8479e-04 - val_loss: 8.2656e-05 Epoch 24/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 6.9879e-05 - val_loss: 6.2360e-05 Epoch 25/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.1980e-04 - val_loss: 4.5659e-05 Epoch 26/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.8073e-05 - val_loss: 4.2629e-05 Epoch 27/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.8982e-05 - val_loss: 3.8475e-05 Epoch 28/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 8.9628e-05 - val_loss: 3.7922e-05 Epoch 29/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.6395e-05 - val_loss: 3.6244e-05 Epoch 30/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 6.0429e-05 - val_loss: 4.7494e-04 Epoch 31/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.7577e-05 - val_loss: 3.4907e-05 Epoch 32/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 4.4244e-05 - val_loss: 4.7000e-05 Epoch 33/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 1.4359e-04 - val_loss: 3.5843e-05 Epoch 34/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 3.5460e-05 - val_loss: 3.6388e-05 Epoch 35/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.6164e-05 - val_loss: 3.6501e-05 Epoch 36/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 8.8671e-05 - val_loss: 3.5552e-05 Epoch 37/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 3.5901e-05 - val_loss: 3.6773e-05 Epoch 38/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 8.2072e-05 - val_loss: 3.5609e-05 Epoch 39/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 3.5580e-05 - val_loss: 3.4665e-05 Epoch 40/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 4.5434e-05 - val_loss: 5.5702e-05 Epoch 41/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.9772e-05 - val_loss: 3.5167e-05 Epoch 42/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 6.9462e-05 - val_loss: 0.0012 Epoch 43/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 9.3966e-05 - val_loss: 3.4895e-05 Epoch 44/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.6265e-05 - val_loss: 5.9290e-05 Epoch 45/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.7843e-05 - val_loss: 3.6437e-05 Epoch 46/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 6.4081e-05 - val_loss: 3.5740e-05 Epoch 47/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 9.0530e-05 - val_loss: 3.4496e-05 Epoch 48/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.6090e-05 - val_loss: 3.7044e-05 Epoch 49/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 8.3593e-05 - val_loss: 3.5292e-05 Epoch 50/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.5681e-05 - val_loss: 3.4150e-05 Epoch 51/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.7302e-05 - val_loss: 3.5704e-05 Epoch 52/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.3372e-05 - val_loss: 3.6931e-05 Epoch 53/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.7722e-05 - val_loss: 3.5659e-05 Epoch 54/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.5542e-05 - val_loss: 4.1661e-05 Epoch 55/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.5800e-05 - val_loss: 3.4699e-05 Epoch 56/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.5152e-05 - val_loss: 3.6531e-05 Epoch 57/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 4.1610e-05 - val_loss: 3.7624e-05 Epoch 58/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 5.3220e-05 - val_loss: 3.2487e-04 Epoch 59/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.4849e-05 - val_loss: 4.0178e-05 Epoch 60/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.2060e-05 - val_loss: 3.6755e-05 Epoch 61/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.8183e-05 - val_loss: 3.6130e-05 Epoch 62/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.3652e-04 - val_loss: 3.8014e-05 Epoch 63/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.5360e-05 - val_loss: 3.4471e-05 Epoch 64/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.5025e-05 - val_loss: 3.5150e-05 Epoch 65/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.2616e-05 - val_loss: 3.6003e-05 Epoch 66/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.9803e-05 - val_loss: 3.8005e-05 Epoch 67/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 4.8341e-05 - val_loss: 1.2148e-04 Epoch 68/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.5740e-05 - val_loss: 3.7297e-05 Epoch 69/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.8722e-05 - val_loss: 3.5301e-05 Epoch 70/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.4143e-05 - val_loss: 3.4342e-05 Epoch 71/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.9729e-05 - val_loss: 3.5244e-05 Epoch 72/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 5.9585e-05 - val_loss: 3.4516e-05 Epoch 73/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 5.4770e-05 - val_loss: 3.6336e-05 Epoch 74/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 3.5945e-05 - val_loss: 3.8451e-05 Epoch 75/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.4603e-05 - val_loss: 3.4487e-05 Epoch 76/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.6313e-05 - val_loss: 3.5674e-05 Epoch 77/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 6.9294e-05 - val_loss: 3.3895e-05 Epoch 78/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.7692e-05 - val_loss: 3.4662e-05 Epoch 79/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 4.0735e-05 - val_loss: 0.0014 Epoch 80/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 7.9858e-05 - val_loss: 3.3427e-05 Epoch 81/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.6114e-05 - val_loss: 3.5275e-05 Epoch 82/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 4.0342e-05 - val_loss: 3.3311e-05 Epoch 83/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.1208e-05 - val_loss: 3.2646e-05 Epoch 84/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.5828e-05 - val_loss: 3.3693e-05 Epoch 85/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.6290e-05 - val_loss: 4.0048e-05 Epoch 86/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 4.6160e-05 - val_loss: 3.4210e-05 Epoch 87/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.0259e-05 - val_loss: 3.4632e-05 Epoch 88/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 6.9006e-05 - val_loss: 2.9006e-04 Epoch 89/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 3.8540e-05 - val_loss: 3.3131e-05 Epoch 90/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.3114e-05 - val_loss: 3.3281e-05 Epoch 91/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 5.1074e-05 - val_loss: 3.1536e-05 Epoch 92/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 3.1837e-05 - val_loss: 3.2542e-05 Epoch 93/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.2386e-05 - val_loss: 3.2199e-05 Epoch 94/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.4039e-05 - val_loss: 1.3513e-04 Epoch 95/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.5875e-05 - val_loss: 3.1827e-05 Epoch 96/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 4.0488e-05 - val_loss: 3.3459e-05 Epoch 97/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.1311e-05 - val_loss: 3.2085e-05 Epoch 98/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.9925e-05 - val_loss: 3.0731e-05 Epoch 99/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.3661e-05 - val_loss: 3.4685e-05 Epoch 100/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.5204e-05 - val_loss: 2.9887e-05 Epoch 101/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.9129e-05 - val_loss: 2.9274e-05 Epoch 102/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 4.2479e-05 - val_loss: 2.7587e-05 Epoch 103/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.7714e-05 - val_loss: 3.0007e-05 Epoch 104/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.2525e-05 - val_loss: 4.2280e-05 Epoch 105/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.6839e-05 - val_loss: 2.7956e-05 Epoch 106/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.0425e-05 - val_loss: 2.5926e-05 Epoch 107/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 2.7498e-05 - val_loss: 2.5288e-05 Epoch 108/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.8758e-05 - val_loss: 2.3848e-05 Epoch 109/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.4998e-05 - val_loss: 2.3406e-05 Epoch 110/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.1984e-05 - val_loss: 2.6967e-05 Epoch 111/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.2264e-05 - val_loss: 2.2215e-05 Epoch 112/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.1111e-05 - val_loss: 2.2008e-05 Epoch 113/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.3385e-05 - val_loss: 6.6078e-05 Epoch 114/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.5740e-05 - val_loss: 1.9789e-05 Epoch 115/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.4469e-05 - val_loss: 2.0019e-05 Epoch 116/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.4200e-05 - val_loss: 4.7974e-05 Epoch 117/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.7359e-05 - val_loss: 1.5383e-05 Epoch 118/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 2.0317e-05 - val_loss: 1.5642e-05 Epoch 119/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.8341e-05 - val_loss: 2.2351e-05 Epoch 120/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 2.5985e-05 - val_loss: 1.3608e-05 Epoch 121/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.4301e-05 - val_loss: 8.1672e-05 Epoch 122/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.8747e-05 - val_loss: 1.2251e-05 Epoch 123/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 1.1870e-05 - val_loss: 1.1693e-05 Epoch 124/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 2.0123e-05 - val_loss: 1.0866e-05 Epoch 125/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.1942e-05 - val_loss: 1.0216e-05 Epoch 126/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 1.3834e-05 - val_loss: 1.0210e-05 Epoch 127/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 3.4737e-05 - val_loss: 5.6461e-04 Epoch 128/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.8131e-05 - val_loss: 8.9750e-06 Epoch 129/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.2382e-05 - val_loss: 8.3610e-06 Epoch 130/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 8.4563e-06 - val_loss: 9.2463e-06 Epoch 131/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 9.7800e-06 - val_loss: 8.2702e-06 Epoch 132/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.2191e-05 - val_loss: 8.2624e-06 Epoch 133/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.6390e-05 - val_loss: 9.5165e-06 Epoch 134/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 8.8670e-06 - val_loss: 7.6018e-06 Epoch 135/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 8.6599e-06 - val_loss: 8.1054e-06 Epoch 136/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.1363e-05 - val_loss: 7.6530e-06 Epoch 137/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.0099e-05 - val_loss: 9.3766e-06 Epoch 138/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.5444e-05 - val_loss: 9.9354e-06 Epoch 139/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.3068e-05 - val_loss: 9.1268e-06 Epoch 140/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.7341e-05 - val_loss: 1.3050e-05 Epoch 141/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.7956e-06 - val_loss: 8.0892e-06 Epoch 142/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 7.4466e-06 - val_loss: 5.9785e-06 Epoch 143/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 7.8923e-06 - val_loss: 5.5913e-06 Epoch 144/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 7.6612e-06 - val_loss: 6.7927e-06 Epoch 145/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.8833e-05 - val_loss: 5.5189e-06 Epoch 146/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 5.7010e-06 - val_loss: 5.5431e-06 Epoch 147/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 7.0593e-06 - val_loss: 5.7793e-06 Epoch 148/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 4.6730e-05 - val_loss: 1.1371e-05 Epoch 149/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.7481e-06 - val_loss: 5.1823e-06 Epoch 150/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.0099e-06 - val_loss: 5.3031e-06 Epoch 151/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 5.3487e-06 - val_loss: 6.5288e-06 Epoch 152/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 5.3411e-06 - val_loss: 5.5283e-06 Epoch 153/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.1459e-06 - val_loss: 5.5785e-06 Epoch 154/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 7.2542e-06 - val_loss: 6.0966e-06 Epoch 155/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.8356e-05 - val_loss: 5.7171e-06 Epoch 156/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.2852e-06 - val_loss: 5.3407e-06 Epoch 157/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 6.0022e-06 - val_loss: 5.3271e-06 Epoch 158/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 7.3480e-06 - val_loss: 5.0394e-06 Epoch 159/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.4610e-06 - val_loss: 1.5985e-05 Epoch 160/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.1606e-05 - val_loss: 2.1315e-05 Epoch 161/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.5169e-05 - val_loss: 5.4833e-06 Epoch 162/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.7518e-06 - val_loss: 6.9106e-06 Epoch 163/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.0344e-05 - val_loss: 5.3913e-06 Epoch 164/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 9.8774e-06 - val_loss: 5.2832e-06 Epoch 165/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.1030e-06 - val_loss: 1.4954e-05 Epoch 166/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 4.1978e-05 - val_loss: 5.0282e-06 Epoch 167/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.0524e-06 - val_loss: 5.8656e-06 Epoch 168/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.7221e-06 - val_loss: 5.0438e-06 Epoch 169/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.1784e-06 - val_loss: 5.9204e-06 Epoch 170/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.5255e-06 - val_loss: 5.8209e-06 Epoch 171/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.2727e-05 - val_loss: 5.1070e-06 Epoch 172/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.5459e-06 - val_loss: 7.6614e-06 Epoch 173/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.1730e-05 - val_loss: 5.6874e-06 Epoch 174/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 7.8309e-06 - val_loss: 4.1466e-05 Epoch 175/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.9474e-05 - val_loss: 1.0766e-05 Epoch 176/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.3732e-06 - val_loss: 7.3699e-06 Epoch 177/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 7.0468e-06 - val_loss: 5.3402e-06 Epoch 178/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.7789e-06 - val_loss: 5.8221e-06 Epoch 179/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.6730e-05 - val_loss: 5.5838e-06 Epoch 180/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.1135e-05 - val_loss: 4.9319e-06 Epoch 181/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.8148e-06 - val_loss: 5.0332e-06 Epoch 182/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.6090e-05 - val_loss: 4.9500e-06 Epoch 183/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.4679e-06 - val_loss: 5.1305e-06 Epoch 184/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 9.0764e-06 - val_loss: 5.1413e-06 Epoch 185/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.8881e-06 - val_loss: 5.0183e-05 Epoch 186/1000 3888/3888 [==============================] - ETA: 0s - loss: 1.9354e-0 - 0s 104us/sample - loss: 1.8154e-05 - val_loss: 5.8104e-06 Epoch 187/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.5169e-06 - val_loss: 5.2057e-06 Epoch 188/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.9678e-05 - val_loss: 5.0411e-06 Epoch 189/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 8.0556e-06 - val_loss: 5.4868e-06 Epoch 190/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.4279e-06 - val_loss: 1.1003e-05 Epoch 191/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.6429e-05 - val_loss: 4.8661e-06 Epoch 192/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.3609e-06 - val_loss: 5.0455e-06 Epoch 193/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.9611e-05 - val_loss: 4.9101e-06 Epoch 194/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.1925e-06 - val_loss: 5.2322e-06 Epoch 195/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.1665e-06 - val_loss: 5.0881e-06 Epoch 196/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.0807e-05 - val_loss: 5.3182e-06 Epoch 197/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.0223e-06 - val_loss: 5.2297e-06 Epoch 198/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 9.3938e-06 - val_loss: 3.5664e-05 Epoch 199/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.2875e-05 - val_loss: 5.3033e-06 Epoch 200/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.9190e-06 - val_loss: 4.9229e-06 Epoch 201/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 9.9934e-06 - val_loss: 5.6191e-05 Epoch 202/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 9.2285e-06 - val_loss: 5.3683e-06 Epoch 203/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.2548e-05 - val_loss: 5.0914e-06 Epoch 204/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.1665e-05 - val_loss: 5.6549e-06 Epoch 205/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.1949e-06 - val_loss: 5.7994e-06 Epoch 206/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.7325e-06 - val_loss: 7.5197e-06 Epoch 207/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.5780e-05 - val_loss: 8.6178e-06 Epoch 208/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.1125e-05 - val_loss: 5.7816e-06 Epoch 209/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.8802e-06 - val_loss: 5.1415e-06 Epoch 210/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 8.7255e-06 - val_loss: 5.9121e-06 Epoch 211/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.1670e-05 - val_loss: 2.0571e-05 Epoch 212/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.4861e-05 - val_loss: 5.3848e-06 Epoch 213/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.3916e-06 - val_loss: 4.8968e-06 Epoch 214/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.6951e-06 - val_loss: 5.0218e-06 Epoch 215/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.8693e-06 - val_loss: 1.4233e-05 Epoch 216/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.5011e-05 - val_loss: 5.5303e-06 Epoch 217/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.1716e-06 - val_loss: 5.3302e-06 Epoch 218/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 8.8016e-06 - val_loss: 5.1528e-06 Epoch 219/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 4.4235e-05 - val_loss: 2.0132e-05 Epoch 220/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 5.3403e-06 - val_loss: 4.9542e-06 Epoch 221/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 5.0615e-06 - val_loss: 4.9248e-06 Epoch 222/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 8.0626e-06 - val_loss: 5.2801e-06 Epoch 223/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 8.1077e-06 - val_loss: 5.0242e-06 Epoch 224/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.1627e-06 - val_loss: 5.1841e-06 Epoch 225/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 8.0086e-06 - val_loss: 5.1013e-06 Epoch 226/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.0874e-05 - val_loss: 4.9499e-06 Epoch 227/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.4019e-06 - val_loss: 4.9033e-06 Epoch 228/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.5019e-05 - val_loss: 1.0005e-05 Epoch 229/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.2218e-06 - val_loss: 5.2911e-06 Epoch 230/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.7086e-06 - val_loss: 5.7231e-06 Epoch 231/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.0339e-05 - val_loss: 5.2158e-06 Epoch 232/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 7.3617e-06 - val_loss: 5.4608e-06 Epoch 233/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.6889e-05 - val_loss: 6.4740e-06 Epoch 234/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.0588e-06 - val_loss: 4.9012e-06 Epoch 235/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.0122e-06 - val_loss: 5.3816e-06 Epoch 236/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.1021e-05 - val_loss: 2.9005e-05 Epoch 237/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.0423e-06 - val_loss: 4.9029e-06 Epoch 238/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.2460e-05 - val_loss: 3.0004e-05 Epoch 239/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.7703e-06 - val_loss: 1.3937e-05 Epoch 240/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.4046e-05 - val_loss: 4.8118e-06 Epoch 241/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.4862e-05 - val_loss: 6.2507e-06 Epoch 242/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.2830e-06 - val_loss: 3.9124e-05 Epoch 243/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.0836e-06 - val_loss: 5.5407e-06 Epoch 244/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.4845e-06 - val_loss: 5.0267e-06 Epoch 245/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.9226e-05 - val_loss: 4.9706e-06 Epoch 246/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.0994e-06 - val_loss: 5.0257e-06 Epoch 247/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.8879e-06 - val_loss: 5.0373e-06 Epoch 248/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.0028e-05 - val_loss: 5.0836e-06 Epoch 249/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 5.0379e-06 - val_loss: 8.0278e-06 Epoch 250/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 9.4061e-06 - val_loss: 5.1660e-06 Epoch 251/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.7664e-05 - val_loss: 4.8208e-06 Epoch 252/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.8100e-06 - val_loss: 5.7327e-06 Epoch 253/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.1340e-06 - val_loss: 4.9193e-06 Epoch 254/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.1229e-06 - val_loss: 1.8027e-05 Epoch 255/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.1558e-05 - val_loss: 3.0126e-05 Epoch 256/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 9.3252e-06 - val_loss: 5.0730e-05 Epoch 257/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 9.1598e-06 - val_loss: 5.0862e-06 Epoch 258/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.2504e-05 - val_loss: 4.7575e-06 Epoch 259/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.8357e-06 - val_loss: 5.1366e-06 Epoch 260/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.3199e-06 - val_loss: 4.7551e-06 Epoch 261/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 5.1272e-06 - val_loss: 5.1383e-06 Epoch 262/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 9.4567e-06 - val_loss: 5.1478e-06 Epoch 263/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.2456e-05 - val_loss: 4.8305e-06 Epoch 264/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.0631e-05 - val_loss: 5.5981e-06 Epoch 265/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 7.5412e-06 - val_loss: 4.8673e-06 Epoch 266/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 8.2073e-06 - val_loss: 4.9963e-06 Epoch 267/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.3912e-05 - val_loss: 3.4565e-05 Epoch 268/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.0892e-05 - val_loss: 5.1535e-06 Epoch 269/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.9878e-06 - val_loss: 1.4471e-05 Epoch 270/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 9.5390e-06 - val_loss: 6.0100e-06 Epoch 271/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.7552e-05 - val_loss: 4.8317e-06 Epoch 272/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.7545e-06 - val_loss: 4.7075e-06 Epoch 273/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.8024e-06 - val_loss: 5.9206e-06 Epoch 274/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.8709e-06 - val_loss: 5.8493e-06 Epoch 275/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.6395e-06 - val_loss: 5.4998e-06 Epoch 276/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.7663e-06 - val_loss: 1.1723e-05 Epoch 277/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 9.0486e-06 - val_loss: 5.1678e-06 Epoch 278/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 1.4844e-05 - val_loss: 4.8489e-06 Epoch 279/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.7815e-06 - val_loss: 4.5568e-06 Epoch 280/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 7.0476e-06 - val_loss: 5.4193e-06 Epoch 281/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.2321e-06 - val_loss: 6.5845e-06 Epoch 282/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.1679e-05 - val_loss: 4.5588e-06 Epoch 283/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.8484e-06 - val_loss: 1.1857e-05 Epoch 284/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.1235e-05 - val_loss: 5.2455e-06 Epoch 285/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.0642e-06 - val_loss: 5.0168e-06 Epoch 286/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.0744e-05 - val_loss: 5.8064e-06 Epoch 287/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.0585e-06 - val_loss: 4.4615e-06 Epoch 288/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.3119e-05 - val_loss: 5.0394e-06 Epoch 289/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.7098e-06 - val_loss: 4.8525e-06 Epoch 290/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 5.1392e-06 - val_loss: 4.9212e-06 Epoch 291/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.4483e-06 - val_loss: 1.6513e-05 Epoch 292/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.1831e-05 - val_loss: 7.4398e-06 Epoch 293/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 7.0615e-06 - val_loss: 9.5160e-05 Epoch 294/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 8.3151e-06 - val_loss: 5.4201e-06 Epoch 295/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.2590e-05 - val_loss: 5.0781e-06 Epoch 296/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.2337e-06 - val_loss: 4.8531e-06 Epoch 297/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 6.3205e-06 - val_loss: 5.0338e-06 Epoch 298/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 5.5627e-06 - val_loss: 5.1430e-06 Epoch 299/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 8.3596e-06 - val_loss: 4.8956e-06 Epoch 300/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.7784e-06 - val_loss: 4.6550e-06 Epoch 301/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 7.9813e-06 - val_loss: 4.9171e-06 Epoch 302/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 9.5712e-06 - val_loss: 5.7721e-06 Epoch 303/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.4205e-06 - val_loss: 3.0211e-05 Epoch 304/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.2759e-05 - val_loss: 4.7967e-06 Epoch 305/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.5602e-06 - val_loss: 4.3862e-06 Epoch 306/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.7119e-06 - val_loss: 4.8089e-06 Epoch 307/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.6186e-06 - val_loss: 4.4769e-06 Epoch 308/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.6707e-06 - val_loss: 4.9086e-06 Epoch 309/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.0940e-05 - val_loss: 3.2448e-05 Epoch 310/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 5.4742e-06 - val_loss: 6.6267e-06 Epoch 311/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 1.4628e-05 - val_loss: 1.3005e-05 Epoch 312/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 4.8071e-06 - val_loss: 4.4598e-06 Epoch 313/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.4748e-05 - val_loss: 4.6374e-06 Epoch 314/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.5457e-06 - val_loss: 4.5341e-06 Epoch 315/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.9733e-05 - val_loss: 3.6640e-05 Epoch 316/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.5945e-06 - val_loss: 4.4687e-06 Epoch 317/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.6621e-06 - val_loss: 5.8700e-06 Epoch 318/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 7.8755e-06 - val_loss: 4.7917e-06 Epoch 319/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.1828e-06 - val_loss: 1.0281e-05 Epoch 320/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 7.8109e-06 - val_loss: 4.4651e-06 Epoch 321/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.9188e-06 - val_loss: 3.5262e-05 Epoch 322/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.7983e-06 - val_loss: 4.9800e-06 Epoch 323/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.5288e-05 - val_loss: 5.2221e-06 Epoch 324/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.2345e-06 - val_loss: 4.5551e-06 Epoch 325/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 7.6294e-06 - val_loss: 4.8709e-06 Epoch 326/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 8.1027e-06 - val_loss: 4.9070e-06 Epoch 327/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.1108e-05 - val_loss: 4.6897e-06 Epoch 328/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.7311e-06 - val_loss: 6.5796e-06 Epoch 329/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.5100e-05 - val_loss: 6.6755e-06 Epoch 330/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.6961e-06 - val_loss: 4.6745e-06 Epoch 331/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.7323e-06 - val_loss: 5.5427e-06 Epoch 332/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.4180e-05 - val_loss: 8.6167e-06 Epoch 333/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 5.3702e-06 - val_loss: 3.9747e-06 Epoch 334/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.2711e-05 - val_loss: 4.1174e-06 Epoch 335/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.5462e-06 - val_loss: 7.7838e-06 Epoch 336/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 7.7333e-06 - val_loss: 4.3035e-06 Epoch 337/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.3112e-05 - val_loss: 4.4719e-06 Epoch 338/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.3166e-05 - val_loss: 5.1118e-06 Epoch 339/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.4038e-06 - val_loss: 4.7201e-06 Epoch 340/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 4.3270e-06 - val_loss: 4.3371e-06 Epoch 341/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.3634e-05 - val_loss: 4.0785e-06 Epoch 342/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.4121e-06 - val_loss: 1.1417e-05 Epoch 343/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.4710e-06 - val_loss: 4.0805e-06 Epoch 344/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 9.5649e-06 - val_loss: 4.1110e-06 Epoch 345/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 4.2755e-06 - val_loss: 4.2250e-06 Epoch 346/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 4.4370e-06 - val_loss: 4.2863e-06 Epoch 347/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.1886e-05 - val_loss: 5.9942e-06 Epoch 348/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 7.2320e-06 - val_loss: 4.3304e-06 Epoch 349/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.8923e-06 - val_loss: 4.5553e-06 Epoch 350/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.8129e-05 - val_loss: 4.2498e-06 Epoch 351/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.9521e-06 - val_loss: 4.3294e-06 Epoch 352/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 7.0268e-06 - val_loss: 8.3489e-06 Epoch 353/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.0720e-06 - val_loss: 4.4433e-06 Epoch 354/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.9607e-06 - val_loss: 3.9563e-06 Epoch 355/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.6156e-06 - val_loss: 4.5936e-06 Epoch 356/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 2.3178e-05 - val_loss: 4.3993e-06 Epoch 357/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.3873e-06 - val_loss: 3.8517e-06 Epoch 358/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.5082e-06 - val_loss: 4.6624e-06 Epoch 359/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 7.4649e-06 - val_loss: 3.9509e-06 Epoch 360/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.7694e-06 - val_loss: 6.2845e-06 Epoch 361/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.2815e-06 - val_loss: 6.3419e-05 Epoch 362/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.9849e-06 - val_loss: 4.6487e-06 Epoch 363/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.0594e-05 - val_loss: 3.7533e-06 Epoch 364/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 1.4823e-05 - val_loss: 3.7482e-06 Epoch 365/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.9819e-06 - val_loss: 4.0929e-06 Epoch 366/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.6992e-05 - val_loss: 3.7075e-06 Epoch 367/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.7237e-06 - val_loss: 3.9358e-06 Epoch 368/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.8160e-06 - val_loss: 4.5256e-06 Epoch 369/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.1589e-06 - val_loss: 3.6337e-06 Epoch 370/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.3154e-05 - val_loss: 4.2920e-06 Epoch 371/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.4231e-06 - val_loss: 3.8406e-06 Epoch 372/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.9863e-06 - val_loss: 4.4476e-06 Epoch 373/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 9.6583e-06 - val_loss: 3.6589e-06 Epoch 374/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.0727e-06 - val_loss: 4.7525e-06 Epoch 375/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.1493e-05 - val_loss: 3.7454e-06 Epoch 376/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.8636e-06 - val_loss: 3.5814e-06 Epoch 377/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.3165e-06 - val_loss: 5.8521e-06 Epoch 378/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.3439e-05 - val_loss: 6.5195e-06 Epoch 379/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.8515e-06 - val_loss: 5.9324e-06 Epoch 380/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 5.3268e-06 - val_loss: 1.0163e-05 Epoch 381/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.0390e-05 - val_loss: 1.2135e-05 Epoch 382/1000 3888/3888 [==============================] - 0s 94us/sample - loss: 6.1042e-06 - val_loss: 3.5871e-06 Epoch 383/1000 3888/3888 [==============================] - 0s 97us/sample - loss: 2.9887e-05 - val_loss: 3.6318e-06 Epoch 384/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.4001e-06 - val_loss: 3.6702e-06 Epoch 385/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.4465e-06 - val_loss: 3.7085e-06 Epoch 386/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.1381e-06 - val_loss: 3.6344e-06 Epoch 387/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.8798e-06 - val_loss: 8.8591e-06 Epoch 388/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 8.1566e-06 - val_loss: 3.4785e-06 Epoch 389/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.0242e-05 - val_loss: 3.8277e-06 Epoch 390/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.4108e-06 - val_loss: 3.4363e-06 Epoch 391/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.6431e-06 - val_loss: 3.5601e-06 Epoch 392/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.2422e-06 - val_loss: 3.6854e-06 Epoch 393/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.3519e-05 - val_loss: 3.3585e-06 Epoch 394/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.8850e-06 - val_loss: 3.5937e-06 Epoch 395/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 6.6792e-06 - val_loss: 3.5633e-06 Epoch 396/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.7257e-05 - val_loss: 1.2816e-05 Epoch 397/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 4.1464e-06 - val_loss: 4.6950e-06 Epoch 398/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.3323e-05 - val_loss: 5.1388e-06 Epoch 399/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.4786e-06 - val_loss: 3.6109e-06 Epoch 400/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.4506e-06 - val_loss: 3.2491e-06 Epoch 401/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.9185e-06 - val_loss: 3.3070e-06 Epoch 402/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 8.1559e-06 - val_loss: 3.3327e-06 Epoch 403/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.8180e-05 - val_loss: 5.2763e-06 Epoch 404/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.7062e-06 - val_loss: 6.2080e-06 Epoch 405/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.6076e-06 - val_loss: 3.2970e-06 Epoch 406/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.0792e-06 - val_loss: 3.0986e-06 Epoch 407/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.5252e-06 - val_loss: 9.7515e-06 Epoch 408/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.8623e-05 - val_loss: 3.0030e-06 Epoch 409/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.1283e-06 - val_loss: 3.3845e-06 Epoch 410/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.5109e-06 - val_loss: 4.4202e-06 Epoch 411/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.1167e-06 - val_loss: 6.5078e-06 Epoch 412/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.9641e-06 - val_loss: 3.2515e-06 Epoch 413/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.7106e-05 - val_loss: 3.1336e-06 Epoch 414/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.4925e-06 - val_loss: 1.4723e-05 Epoch 415/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.0289e-06 - val_loss: 5.8728e-06 Epoch 416/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 4.1501e-06 - val_loss: 3.8416e-06 Epoch 417/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 7.5721e-06 - val_loss: 3.2010e-06 Epoch 418/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.2196e-06 - val_loss: 4.1598e-06 Epoch 419/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.0746e-05 - val_loss: 3.4600e-06 Epoch 420/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.2259e-06 - val_loss: 2.9926e-06 Epoch 421/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.4002e-06 - val_loss: 3.2789e-06 Epoch 422/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.1587e-06 - val_loss: 3.0755e-06 Epoch 423/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 8.1442e-06 - val_loss: 3.6701e-06 Epoch 424/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.0343e-05 - val_loss: 3.1752e-06 Epoch 425/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.1246e-06 - val_loss: 5.6706e-06 Epoch 426/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.8480e-06 - val_loss: 3.0396e-06 Epoch 427/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.6480e-06 - val_loss: 5.6841e-06 Epoch 428/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.0143e-05 - val_loss: 4.2139e-06 Epoch 429/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 7.6166e-06 - val_loss: 5.0772e-06 Epoch 430/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.0705e-05 - val_loss: 3.1488e-06 Epoch 431/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.8952e-06 - val_loss: 3.7206e-06 Epoch 432/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.0225e-05 - val_loss: 3.1733e-06 Epoch 433/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.1882e-06 - val_loss: 2.8599e-06 Epoch 434/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.6224e-06 - val_loss: 3.8469e-06 Epoch 435/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 9.3014e-06 - val_loss: 4.2564e-05 Epoch 436/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 7.3309e-06 - val_loss: 3.3455e-06 Epoch 437/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.2192e-06 - val_loss: 3.4716e-06 Epoch 438/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.2653e-05 - val_loss: 2.5595e-05 Epoch 439/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.7822e-06 - val_loss: 3.9752e-06 Epoch 440/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.7849e-06 - val_loss: 3.1251e-06 Epoch 441/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.5016e-06 - val_loss: 5.4101e-06 Epoch 442/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.3127e-05 - val_loss: 3.2529e-06 Epoch 443/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 8.1172e-06 - val_loss: 4.0893e-06 Epoch 444/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.4186e-05 - val_loss: 2.9175e-06 Epoch 445/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 4.8729e-06 - val_loss: 1.9069e-05 Epoch 446/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.4402e-06 - val_loss: 3.1424e-06 Epoch 447/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.4015e-06 - val_loss: 2.7400e-06 Epoch 448/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.3899e-05 - val_loss: 4.7651e-06 Epoch 449/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.0546e-06 - val_loss: 2.9096e-06 Epoch 450/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.4416e-06 - val_loss: 9.3531e-06 Epoch 451/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.0130e-06 - val_loss: 2.8515e-06 Epoch 452/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.1875e-06 - val_loss: 3.8936e-06 Epoch 453/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 9.3976e-06 - val_loss: 2.7165e-06 Epoch 454/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.6241e-06 - val_loss: 3.0325e-05 Epoch 455/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 8.4714e-06 - val_loss: 3.1062e-06 Epoch 456/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.8800e-06 - val_loss: 7.4775e-06 Epoch 457/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.1552e-05 - val_loss: 3.0370e-06 Epoch 458/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.7559e-06 - val_loss: 2.9865e-06 Epoch 459/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 9.4194e-06 - val_loss: 3.4532e-06 Epoch 460/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.8423e-06 - val_loss: 2.8869e-06 Epoch 461/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.5850e-06 - val_loss: 2.9619e-06 Epoch 462/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 8.0454e-06 - val_loss: 2.6364e-06 Epoch 463/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.5436e-06 - val_loss: 8.0159e-06 Epoch 464/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.7730e-05 - val_loss: 5.0846e-06 Epoch 465/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.6755e-06 - val_loss: 2.8201e-06 Epoch 466/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 5.6932e-06 - val_loss: 3.1177e-06 Epoch 467/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 7.1501e-06 - val_loss: 3.7080e-06 Epoch 468/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 3.0300e-06 - val_loss: 2.5705e-06 Epoch 469/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 3.0099e-06 - val_loss: 3.0578e-06 Epoch 470/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.9435e-06 - val_loss: 3.6098e-05 Epoch 471/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.2938e-05 - val_loss: 2.6384e-06 Epoch 472/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.4135e-06 - val_loss: 2.6095e-06 Epoch 473/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.1528e-06 - val_loss: 3.1163e-06 Epoch 474/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.5405e-05 - val_loss: 2.7423e-06 Epoch 475/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.7039e-06 - val_loss: 2.5798e-06 Epoch 476/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.6428e-06 - val_loss: 2.8517e-06 Epoch 477/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 9.6302e-06 - val_loss: 4.5265e-06 Epoch 478/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.3244e-06 - val_loss: 3.0128e-06 Epoch 479/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 6.3652e-06 - val_loss: 2.8536e-06 Epoch 480/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.8021e-06 - val_loss: 2.9453e-06 Epoch 481/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 9.1304e-06 - val_loss: 1.0012e-05 Epoch 482/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.2842e-06 - val_loss: 3.4515e-06 Epoch 483/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 5.1865e-06 - val_loss: 7.6181e-06 Epoch 484/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 4.8024e-06 - val_loss: 2.6804e-06 Epoch 485/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.5376e-05 - val_loss: 2.5777e-06 Epoch 486/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.8003e-06 - val_loss: 2.8430e-06 Epoch 487/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 9.6118e-06 - val_loss: 3.4135e-06 Epoch 488/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.9655e-06 - val_loss: 2.6527e-06 Epoch 489/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.9998e-06 - val_loss: 2.8318e-06 Epoch 490/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 5.0941e-06 - val_loss: 5.0448e-06 Epoch 491/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.1511e-05 - val_loss: 3.2270e-06 Epoch 492/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.8192e-06 - val_loss: 2.7097e-06 Epoch 493/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.6374e-06 - val_loss: 2.4748e-06 Epoch 494/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.1791e-06 - val_loss: 2.8073e-06 Epoch 495/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 7.0822e-06 - val_loss: 2.7407e-06 Epoch 496/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.5561e-06 - val_loss: 2.6496e-06 Epoch 497/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.1056e-05 - val_loss: 7.5975e-06 Epoch 498/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.6853e-06 - val_loss: 2.8118e-06 Epoch 499/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.5395e-05 - val_loss: 2.7285e-06 Epoch 500/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.6706e-06 - val_loss: 2.4853e-06 Epoch 501/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.2579e-06 - val_loss: 2.4449e-06 Epoch 502/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.1775e-06 - val_loss: 2.7373e-06 Epoch 503/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 3.2772e-06 - val_loss: 2.8525e-06 Epoch 504/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 6.4795e-06 - val_loss: 1.2114e-05 Epoch 505/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 4.1784e-06 - val_loss: 2.4157e-06 Epoch 506/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.4070e-06 - val_loss: 2.8884e-06 Epoch 507/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 8.4590e-06 - val_loss: 2.4394e-06 Epoch 508/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.9955e-06 - val_loss: 2.4539e-06 Epoch 509/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.6140e-05 - val_loss: 1.7622e-04 Epoch 510/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 7.1826e-06 - val_loss: 2.6850e-06 Epoch 511/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.6511e-06 - val_loss: 2.6438e-06 Epoch 512/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.3145e-06 - val_loss: 2.6619e-06 Epoch 513/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.6595e-06 - val_loss: 3.0094e-06 Epoch 514/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.7332e-05 - val_loss: 2.8006e-06 Epoch 515/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.5385e-06 - val_loss: 3.1653e-06 Epoch 516/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.7602e-06 - val_loss: 2.7518e-06 Epoch 517/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 3.7818e-06 - val_loss: 3.2569e-06 Epoch 518/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.3988e-06 - val_loss: 2.5112e-06 Epoch 519/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.0732e-06 - val_loss: 2.6304e-06 Epoch 520/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 9.0399e-06 - val_loss: 2.6416e-06 Epoch 521/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.1050e-06 - val_loss: 2.8422e-06 Epoch 522/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 3.0254e-05 - val_loss: 2.4614e-06 Epoch 523/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.3949e-06 - val_loss: 2.3782e-06 Epoch 524/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.4348e-06 - val_loss: 2.4890e-06 Epoch 525/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.0332e-06 - val_loss: 3.2098e-05 Epoch 526/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.8127e-06 - val_loss: 2.6270e-06 Epoch 527/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.5216e-06 - val_loss: 2.6958e-06 Epoch 528/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.8618e-06 - val_loss: 2.9628e-06 Epoch 529/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.6415e-06 - val_loss: 3.1317e-06 Epoch 530/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.0715e-06 - val_loss: 1.6132e-05 Epoch 531/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 8.1260e-06 - val_loss: 1.5898e-04 Epoch 532/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.0176e-05 - val_loss: 2.2813e-06 Epoch 533/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.4526e-06 - val_loss: 2.6152e-06 Epoch 534/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.4169e-06 - val_loss: 3.2822e-06 Epoch 535/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.2861e-06 - val_loss: 2.4041e-06 Epoch 536/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.9049e-06 - val_loss: 3.2604e-06 Epoch 537/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 9.4201e-06 - val_loss: 2.7821e-06 Epoch 538/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.5130e-06 - val_loss: 7.6638e-06 Epoch 539/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.9377e-06 - val_loss: 2.9142e-06 Epoch 540/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 3.9291e-06 - val_loss: 2.6368e-06 Epoch 541/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.7780e-06 - val_loss: 1.5474e-04 Epoch 542/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.4491e-05 - val_loss: 2.4348e-06 Epoch 543/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.5306e-06 - val_loss: 2.6341e-06 Epoch 544/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 7.2642e-06 - val_loss: 2.5398e-06 Epoch 545/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 2.4919e-06 - val_loss: 3.2287e-06 Epoch 546/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.2073e-06 - val_loss: 1.1846e-05 Epoch 547/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.0164e-05 - val_loss: 2.4015e-06 Epoch 548/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.8959e-06 - val_loss: 3.4314e-06 Epoch 549/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.8666e-06 - val_loss: 2.7536e-06 Epoch 550/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 9.3627e-06 - val_loss: 2.3535e-06 Epoch 551/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.3647e-06 - val_loss: 2.2738e-06 Epoch 552/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.0272e-06 - val_loss: 2.3279e-06 Epoch 553/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.0694e-05 - val_loss: 2.8632e-06 Epoch 554/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.4053e-06 - val_loss: 7.4176e-06 Epoch 555/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 8.9225e-06 - val_loss: 2.2161e-06 Epoch 556/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.2146e-06 - val_loss: 2.3061e-06 Epoch 557/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.8858e-06 - val_loss: 3.3058e-06 Epoch 558/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.7328e-06 - val_loss: 3.8837e-06 Epoch 559/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.4267e-06 - val_loss: 2.0716e-06 Epoch 560/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.5461e-06 - val_loss: 9.9572e-06 Epoch 561/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 8.2315e-06 - val_loss: 2.1343e-06 Epoch 562/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 7.6250e-06 - val_loss: 2.1587e-06 Epoch 563/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.8253e-06 - val_loss: 2.1548e-06 Epoch 564/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.0960e-06 - val_loss: 2.7204e-06 Epoch 565/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 3.4280e-06 - val_loss: 1.0212e-05 Epoch 566/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.1501e-05 - val_loss: 2.2023e-06 Epoch 567/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.9056e-06 - val_loss: 2.1986e-06 Epoch 568/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 8.9194e-06 - val_loss: 2.1831e-06 Epoch 569/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.3222e-06 - val_loss: 2.0476e-06 Epoch 570/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.3505e-06 - val_loss: 2.1675e-06 Epoch 571/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 3.4766e-06 - val_loss: 2.4266e-06 Epoch 572/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.0426e-05 - val_loss: 2.5058e-06 Epoch 573/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 5.7065e-06 - val_loss: 6.6395e-06 Epoch 574/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.1668e-06 - val_loss: 3.0820e-06 Epoch 575/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.9130e-06 - val_loss: 2.0596e-06 Epoch 576/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 5.2683e-06 - val_loss: 6.1442e-06 Epoch 577/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 5.5059e-06 - val_loss: 2.1132e-06 Epoch 578/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.9460e-06 - val_loss: 1.9898e-06 Epoch 579/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 8.2428e-06 - val_loss: 2.1241e-06 Epoch 580/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.7587e-06 - val_loss: 2.2469e-06 Epoch 581/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 9.8901e-06 - val_loss: 2.0694e-06 Epoch 582/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.3967e-06 - val_loss: 1.9195e-06 Epoch 583/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.1953e-06 - val_loss: 2.1298e-06 Epoch 584/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.9616e-06 - val_loss: 2.4264e-06 Epoch 585/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.3160e-06 - val_loss: 2.9170e-06 Epoch 586/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.6424e-06 - val_loss: 2.1540e-06 Epoch 587/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.2708e-06 - val_loss: 1.9728e-04 Epoch 588/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 7.6577e-06 - val_loss: 1.9258e-06 Epoch 589/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.8769e-06 - val_loss: 2.5844e-06 Epoch 590/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.1559e-06 - val_loss: 2.2710e-06 Epoch 591/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.2033e-06 - val_loss: 2.4007e-06 Epoch 592/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.2184e-06 - val_loss: 2.5670e-05 Epoch 593/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.4938e-06 - val_loss: 2.3584e-06 Epoch 594/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.8161e-06 - val_loss: 3.0545e-06 Epoch 595/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.5915e-06 - val_loss: 1.2822e-05 Epoch 596/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.2210e-06 - val_loss: 3.1632e-06 Epoch 597/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.9176e-06 - val_loss: 2.1339e-06 Epoch 598/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.6785e-06 - val_loss: 3.1807e-06 Epoch 599/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.3373e-06 - val_loss: 1.4535e-05 Epoch 600/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.7131e-06 - val_loss: 1.7612e-06 Epoch 601/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.6269e-06 - val_loss: 3.0432e-06 Epoch 602/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.0552e-05 - val_loss: 1.8769e-06 Epoch 603/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.5103e-06 - val_loss: 2.0281e-06 Epoch 604/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.0693e-05 - val_loss: 2.0279e-06 Epoch 605/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.0089e-06 - val_loss: 1.9233e-06 Epoch 606/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.7832e-06 - val_loss: 1.7517e-06 Epoch 607/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.9150e-06 - val_loss: 3.2185e-06 Epoch 608/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 8.2165e-06 - val_loss: 2.7827e-06 Epoch 609/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.1336e-06 - val_loss: 1.6807e-06 Epoch 610/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.0444e-06 - val_loss: 1.9298e-06 Epoch 611/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 4.6912e-06 - val_loss: 2.3497e-06 Epoch 612/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 4.3254e-06 - val_loss: 4.1531e-06 Epoch 613/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.3826e-05 - val_loss: 1.5368e-04 Epoch 614/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.7142e-06 - val_loss: 1.8502e-06 Epoch 615/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.5239e-06 - val_loss: 2.0249e-06 Epoch 616/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.0531e-06 - val_loss: 2.0639e-06 Epoch 617/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.9105e-06 - val_loss: 2.9217e-06 Epoch 618/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 8.1239e-06 - val_loss: 1.8108e-06 Epoch 619/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.4310e-06 - val_loss: 1.9587e-06 Epoch 620/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.7611e-06 - val_loss: 1.5661e-06 Epoch 621/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.1052e-06 - val_loss: 3.2501e-06 Epoch 622/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 7.3067e-06 - val_loss: 2.9066e-05 Epoch 623/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.4670e-06 - val_loss: 1.6921e-06 Epoch 624/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.3248e-06 - val_loss: 1.6399e-06 Epoch 625/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 4.2929e-06 - val_loss: 1.5643e-06 Epoch 626/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.0256e-05 - val_loss: 1.6999e-06 Epoch 627/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.9151e-06 - val_loss: 1.6844e-06 Epoch 628/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.6571e-06 - val_loss: 1.7718e-06 Epoch 629/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.6319e-06 - val_loss: 2.9774e-06 Epoch 630/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.1562e-06 - val_loss: 3.1268e-06 Epoch 631/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.0154e-06 - val_loss: 1.6568e-06 Epoch 632/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 9.9970e-06 - val_loss: 3.3753e-06 Epoch 633/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.0242e-06 - val_loss: 1.4710e-06 Epoch 634/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.9737e-06 - val_loss: 5.0970e-06 Epoch 635/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.7710e-06 - val_loss: 1.9928e-06 Epoch 636/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.9441e-06 - val_loss: 1.5312e-06 Epoch 637/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.8332e-06 - val_loss: 1.5876e-06 Epoch 638/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.0395e-05 - val_loss: 1.4889e-06 Epoch 639/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.8108e-06 - val_loss: 9.1968e-06 Epoch 640/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.2025e-06 - val_loss: 2.6203e-06 Epoch 641/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.5283e-06 - val_loss: 1.5452e-06 Epoch 642/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.6146e-06 - val_loss: 3.4278e-06 Epoch 643/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.7904e-06 - val_loss: 1.4155e-06 Epoch 644/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 1.1134e-05 - val_loss: 1.5511e-06 Epoch 645/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.8921e-06 - val_loss: 1.5820e-06 Epoch 646/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.7583e-06 - val_loss: 1.7527e-06 Epoch 647/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.2230e-05 - val_loss: 1.1869e-05 Epoch 648/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.9719e-06 - val_loss: 1.6037e-06 Epoch 649/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.6896e-06 - val_loss: 1.6736e-06 Epoch 650/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.8430e-06 - val_loss: 1.3596e-06 Epoch 651/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.6486e-06 - val_loss: 3.9974e-06 Epoch 652/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.7020e-06 - val_loss: 2.4310e-06 Epoch 653/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.6078e-06 - val_loss: 1.4435e-06 Epoch 654/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.0852e-06 - val_loss: 1.4949e-06 Epoch 655/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 8.4255e-06 - val_loss: 1.8559e-05 Epoch 656/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.7205e-06 - val_loss: 1.3922e-06 Epoch 657/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.7953e-06 - val_loss: 2.4881e-06 Epoch 658/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.5690e-06 - val_loss: 1.5030e-06 Epoch 659/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.4997e-06 - val_loss: 1.5249e-06 Epoch 660/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.9972e-06 - val_loss: 2.1769e-06 Epoch 661/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.9981e-06 - val_loss: 4.6937e-06 Epoch 662/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.6219e-05 - val_loss: 2.3201e-06 Epoch 663/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.6286e-06 - val_loss: 1.4623e-06 Epoch 664/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.6818e-06 - val_loss: 9.8746e-06 Epoch 665/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.9951e-06 - val_loss: 2.0627e-06 Epoch 666/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 9.3081e-06 - val_loss: 8.1998e-06 Epoch 667/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.5016e-06 - val_loss: 2.1251e-06 Epoch 668/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.2423e-06 - val_loss: 3.9096e-06 Epoch 669/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.7451e-06 - val_loss: 1.5208e-06 Epoch 670/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.5929e-06 - val_loss: 1.7533e-06 Epoch 671/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.8447e-06 - val_loss: 5.1322e-05 Epoch 672/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 7.2189e-06 - val_loss: 1.9987e-06 Epoch 673/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.4331e-06 - val_loss: 1.6382e-06 Epoch 674/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.0510e-05 - val_loss: 1.7097e-05 Epoch 675/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.9443e-06 - val_loss: 1.5095e-06 Epoch 676/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.7940e-06 - val_loss: 1.9269e-06 Epoch 677/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.1738e-06 - val_loss: 1.5253e-06 Epoch 678/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 5.8129e-06 - val_loss: 3.6360e-06 Epoch 679/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.0350e-06 - val_loss: 1.4185e-06 Epoch 680/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.8244e-06 - val_loss: 2.9612e-06 Epoch 681/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.6790e-06 - val_loss: 5.9317e-06 Epoch 682/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.8968e-06 - val_loss: 1.3753e-06 Epoch 683/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 7.0388e-06 - val_loss: 1.2901e-05 Epoch 684/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.5089e-06 - val_loss: 1.5053e-06 Epoch 685/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.7561e-06 - val_loss: 1.9657e-05 Epoch 686/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.3843e-06 - val_loss: 1.5935e-06 Epoch 687/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.8617e-06 - val_loss: 1.2060e-06 Epoch 688/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.2219e-06 - val_loss: 6.3618e-06 Epoch 689/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 8.8605e-06 - val_loss: 1.3222e-06 Epoch 690/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.3331e-06 - val_loss: 3.3300e-06 Epoch 691/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.6368e-06 - val_loss: 1.3297e-06 Epoch 692/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.8329e-06 - val_loss: 2.3770e-06 Epoch 693/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.8043e-06 - val_loss: 2.5965e-06 Epoch 694/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.9277e-06 - val_loss: 1.6307e-06 Epoch 695/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.6013e-06 - val_loss: 5.7026e-05 Epoch 696/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.3196e-06 - val_loss: 3.7107e-05 Epoch 697/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.9436e-06 - val_loss: 1.6274e-06 Epoch 698/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 9.2553e-06 - val_loss: 1.0876e-06 Epoch 699/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.6584e-06 - val_loss: 1.1857e-06 Epoch 700/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.2122e-06 - val_loss: 2.7331e-06 Epoch 701/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.1122e-05 - val_loss: 1.3649e-06 Epoch 702/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.4372e-06 - val_loss: 3.6740e-06 Epoch 703/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 4.5740e-06 - val_loss: 1.4361e-06 Epoch 704/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 1.3835e-06 - val_loss: 1.1924e-06 Epoch 705/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.9760e-06 - val_loss: 1.3695e-06 Epoch 706/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.7700e-06 - val_loss: 1.0775e-06 Epoch 707/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.1717e-06 - val_loss: 2.0667e-06 Epoch 708/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.3456e-06 - val_loss: 1.5766e-06 Epoch 709/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.9796e-06 - val_loss: 7.6243e-06 Epoch 710/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 4.3017e-06 - val_loss: 1.3453e-06 Epoch 711/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.0637e-05 - val_loss: 1.0411e-06 Epoch 712/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 1.5320e-06 - val_loss: 2.1798e-06 Epoch 713/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 1.8862e-06 - val_loss: 3.4081e-06 Epoch 714/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 6.3535e-06 - val_loss: 1.0466e-06 Epoch 715/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.3345e-06 - val_loss: 4.9182e-06 Epoch 716/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.7764e-06 - val_loss: 8.9795e-06 Epoch 717/1000 3888/3888 [==============================] - 0s 98us/sample - loss: 6.4241e-06 - val_loss: 1.9376e-06 Epoch 718/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.7059e-06 - val_loss: 2.8663e-06 Epoch 719/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.0772e-06 - val_loss: 1.0808e-05 Epoch 720/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.8220e-06 - val_loss: 1.1734e-06 Epoch 721/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.3519e-06 - val_loss: 1.1421e-06 Epoch 722/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.0362e-06 - val_loss: 6.7849e-06 Epoch 723/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.3880e-06 - val_loss: 1.1539e-06 Epoch 724/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.8311e-06 - val_loss: 1.6307e-06 Epoch 725/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.5614e-06 - val_loss: 1.5835e-06 Epoch 726/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.7161e-06 - val_loss: 1.4577e-06 Epoch 727/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.7547e-06 - val_loss: 1.2287e-06 Epoch 728/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.0986e-06 - val_loss: 1.0419e-06 Epoch 729/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.4966e-06 - val_loss: 1.0001e-06 Epoch 730/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.2007e-06 - val_loss: 1.4794e-05 Epoch 731/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.4366e-06 - val_loss: 1.0259e-06 Epoch 732/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.8042e-06 - val_loss: 6.0819e-06 Epoch 733/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 6.6389e-06 - val_loss: 2.2165e-06 Epoch 734/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.6723e-06 - val_loss: 2.3272e-06 Epoch 735/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 3.6584e-06 - val_loss: 1.0546e-06 Epoch 736/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.9420e-06 - val_loss: 1.0166e-06 Epoch 737/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.1281e-06 - val_loss: 1.2723e-06 Epoch 738/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 8.9904e-06 - val_loss: 1.2738e-06 Epoch 739/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.2107e-06 - val_loss: 1.0656e-06 Epoch 740/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 6.3808e-06 - val_loss: 2.1759e-06 Epoch 741/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.4318e-06 - val_loss: 1.2610e-06 Epoch 742/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.5476e-06 - val_loss: 2.1355e-06 Epoch 743/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 5.5866e-06 - val_loss: 3.5372e-06 Epoch 744/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.9487e-06 - val_loss: 1.9419e-06 Epoch 745/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.2192e-06 - val_loss: 1.6982e-06 Epoch 746/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.3756e-06 - val_loss: 1.6649e-06 Epoch 747/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.0350e-06 - val_loss: 1.2832e-06 Epoch 748/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.5300e-05 - val_loss: 1.1413e-06 Epoch 749/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.7986e-06 - val_loss: 7.6191e-06 Epoch 750/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.3629e-06 - val_loss: 5.3663e-06 Epoch 751/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.9229e-06 - val_loss: 1.5807e-06 Epoch 752/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.1316e-06 - val_loss: 2.5694e-06 Epoch 753/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.5875e-06 - val_loss: 1.6416e-06 Epoch 754/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 5.6695e-06 - val_loss: 1.1779e-06 Epoch 755/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.9264e-06 - val_loss: 1.4852e-06 Epoch 756/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.1417e-06 - val_loss: 4.0825e-06 Epoch 757/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.9968e-06 - val_loss: 1.8138e-06 Epoch 758/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.2203e-06 - val_loss: 5.1587e-06 Epoch 759/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 3.7198e-06 - val_loss: 9.8413e-07 Epoch 760/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 5.9857e-06 - val_loss: 2.3127e-06 Epoch 761/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.7601e-06 - val_loss: 9.1543e-07 Epoch 762/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 7.0408e-06 - val_loss: 9.9067e-07 Epoch 763/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.8405e-06 - val_loss: 1.3683e-06 Epoch 764/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.0318e-06 - val_loss: 3.1831e-06 Epoch 765/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 7.4205e-06 - val_loss: 5.4743e-06 Epoch 766/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.1470e-06 - val_loss: 2.2197e-06 Epoch 767/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 2.8162e-06 - val_loss: 1.3075e-06 Epoch 768/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.3813e-06 - val_loss: 1.1984e-06 Epoch 769/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.1976e-06 - val_loss: 1.1190e-06 Epoch 770/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.7294e-06 - val_loss: 7.5016e-06 Epoch 771/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 3.1380e-06 - val_loss: 1.0088e-06 Epoch 772/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 9.6019e-06 - val_loss: 1.0890e-06 Epoch 773/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 2.6043e-06 - val_loss: 1.1233e-06 Epoch 774/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.7455e-06 - val_loss: 1.5507e-06 Epoch 775/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.7537e-06 - val_loss: 2.4299e-06 Epoch 776/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 6.9178e-06 - val_loss: 1.0166e-06 Epoch 777/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 2.2422e-06 - val_loss: 4.7166e-06 Epoch 778/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.3663e-06 - val_loss: 1.0086e-06 Epoch 779/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.8729e-06 - val_loss: 1.6734e-06 Epoch 780/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.6128e-06 - val_loss: 3.1822e-06 Epoch 781/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 1.9689e-06 - val_loss: 1.1789e-05 Epoch 782/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 4.5305e-06 - val_loss: 1.0456e-06 Epoch 783/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 3.5889e-06 - val_loss: 1.3464e-06 Epoch 784/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.8147e-06 - val_loss: 1.0526e-06 Epoch 785/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.6319e-06 - val_loss: 1.1504e-06 Epoch 786/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.4002e-06 - val_loss: 1.2570e-06 Epoch 787/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.4318e-05 - val_loss: 1.1011e-06 Epoch 788/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.0569e-06 - val_loss: 1.2097e-06 Epoch 789/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.5918e-06 - val_loss: 1.0921e-06 Epoch 790/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.4786e-06 - val_loss: 1.1854e-05 Epoch 791/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 4.3297e-06 - val_loss: 1.0324e-06 Epoch 792/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.7167e-06 - val_loss: 1.1346e-05 Epoch 793/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 1.9370e-06 - val_loss: 1.2745e-06 Epoch 794/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.7781e-06 - val_loss: 2.3013e-06 Epoch 795/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 1.3460e-06 - val_loss: 1.2996e-06 Epoch 796/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 2.2860e-06 - val_loss: 7.4163e-06 Epoch 797/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 8.1239e-06 - val_loss: 1.4879e-05 Epoch 798/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.5468e-06 - val_loss: 1.0941e-06 Epoch 799/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 2.1892e-06 - val_loss: 6.7516e-06 Epoch 800/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 7.4627e-06 - val_loss: 1.2901e-06 Epoch 801/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 2.0532e-06 - val_loss: 1.4233e-05 Epoch 802/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 6.7566e-06 - val_loss: 1.5241e-06 Epoch 803/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.2775e-06 - val_loss: 1.0330e-06 Epoch 804/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.2983e-06 - val_loss: 3.0350e-06 Epoch 805/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.8705e-06 - val_loss: 1.9995e-06 Epoch 806/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 2.1710e-06 - val_loss: 4.7666e-06 Epoch 807/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 5.2751e-06 - val_loss: 1.3372e-06 Epoch 808/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.9725e-06 - val_loss: 1.3906e-06 Epoch 809/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.0348e-06 - val_loss: 4.1727e-06 Epoch 810/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.1942e-06 - val_loss: 9.9325e-07 Epoch 811/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 1.8222e-06 - val_loss: 1.2197e-06 Epoch 812/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 9.1813e-06 - val_loss: 1.3247e-06 Epoch 813/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.4130e-06 - val_loss: 1.0631e-06 Epoch 814/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.1664e-06 - val_loss: 1.1733e-06 Epoch 815/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 3.6440e-06 - val_loss: 1.4741e-06 Epoch 816/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.5945e-06 - val_loss: 2.5588e-06 Epoch 817/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 3.8969e-06 - val_loss: 5.0881e-06 Epoch 818/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.0909e-06 - val_loss: 1.2971e-05 Epoch 819/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 3.6708e-06 - val_loss: 1.1811e-06 Epoch 820/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 9.0817e-06 - val_loss: 9.7526e-07 Epoch 821/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.1499e-06 - val_loss: 4.6765e-05 Epoch 822/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.9781e-06 - val_loss: 3.1292e-06 Epoch 823/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.9257e-06 - val_loss: 1.1358e-06 Epoch 824/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.6832e-06 - val_loss: 2.2892e-06 Epoch 825/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.0017e-05 - val_loss: 9.8985e-07 Epoch 826/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.9602e-06 - val_loss: 9.7654e-07 Epoch 827/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.5129e-06 - val_loss: 2.7151e-06 Epoch 828/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.5211e-06 - val_loss: 3.0591e-06 Epoch 829/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.1787e-06 - val_loss: 5.4666e-05 Epoch 830/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.0702e-06 - val_loss: 3.6713e-06 Epoch 831/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 1.9802e-06 - val_loss: 2.7535e-06 Epoch 832/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 3.9714e-06 - val_loss: 1.0601e-06 Epoch 833/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 2.0874e-06 - val_loss: 1.4974e-06 Epoch 834/1000 3888/3888 [==============================] - 0s 100us/sample - loss: 4.6706e-06 - val_loss: 2.1163e-06 Epoch 835/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 6.9537e-06 - val_loss: 1.2902e-06 Epoch 836/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.0588e-06 - val_loss: 2.5100e-06 Epoch 837/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 5.0577e-06 - val_loss: 1.1658e-06 Epoch 838/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 2.1714e-06 - val_loss: 1.0455e-06 Epoch 839/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 8.7025e-06 - val_loss: 1.9254e-05 Epoch 840/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.8381e-06 - val_loss: 2.4443e-06 Epoch 841/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 2.7220e-06 - val_loss: 1.4713e-06 Epoch 842/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 4.3758e-06 - val_loss: 9.2539e-07 Epoch 843/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.7809e-06 - val_loss: 7.9680e-06 Epoch 844/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 4.0015e-06 - val_loss: 4.2171e-06 Epoch 845/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 5.8430e-06 - val_loss: 2.2721e-05 Epoch 846/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 6.0770e-06 - val_loss: 1.0480e-06 Epoch 847/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.2198e-06 - val_loss: 1.3538e-06 Epoch 848/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.1428e-06 - val_loss: 1.3783e-06 Epoch 849/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 9.0912e-06 - val_loss: 2.5423e-06 Epoch 850/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 2.3514e-06 - val_loss: 6.5171e-06 Epoch 851/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 7.6661e-06 - val_loss: 3.4803e-06 Epoch 852/1000 3888/3888 [==============================] - 0s 102us/sample - loss: 1.8125e-06 - val_loss: 9.5794e-07 Epoch 853/1000 3888/3888 [==============================] - 0s 103us/sample - loss: 1.7280e-06 - val_loss: 1.3860e-06 Epoch 854/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.3451e-06 - val_loss: 1.4447e-06 Epoch 855/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 3.2464e-06 - val_loss: 1.7784e-06 Epoch 856/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 3.2174e-06 - val_loss: 9.8018e-06 Epoch 857/1000 3888/3888 [==============================] - ETA: 0s - loss: 4.7744e-0 - 0s 105us/sample - loss: 4.0797e-06 - val_loss: 1.1600e-06 Epoch 858/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 5.0555e-06 - val_loss: 3.4396e-06 Epoch 859/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 6.4839e-06 - val_loss: 2.0227e-05 Epoch 860/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 1.6769e-06 - val_loss: 1.4328e-05 Epoch 861/1000 3200/3888 [=======================>......] - ETA: 0s - loss: 3.2997e-06Restoring model weights from the end of the best epoch. 3888/3888 [==============================] - 0s 103us/sample - loss: 3.0834e-06 - val_loss: 6.3939e-06 Epoch 00861: early stopping
In [32]:
print(history.history.keys())
print('best value: ', autoencoder.evaluate(X_train_1D_norm, X_train_1D_norm, verbose=0))
pd.DataFrame(history.history).plot(figsize=(8, 5), logy=True)
plt.grid()
dict_keys(['loss', 'val_loss']) best value: 9.154292058142889e-07
In [33]:
X_reconstructions = autoencoder.predict(X_train_1D_norm)
X_reconstructions = stdscaler.inverse_transform(X_reconstructions)
In [34]:
calculateerror(X_train_1D.reshape(len(times),len(groups),nl,nc),
X_reconstructions.reshape(len(times),len(groups),nl,nc),
groups,
print_step=0)
max_abs_error: 14.3992919921875 mean_abs_error: 0.02822514491985846
/home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: divide by zero encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until
In [35]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(X_reconstructions.reshape(len(times),len(groups),nl,nc)[100,i,:,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [36]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_train_1D[:,i*nl*nc+4])
ax.flatten()[i].plot(times, X_reconstructions[:,i*nl*nc+4],'--')
ax.flatten()[i].set_title(group)
In [37]:
np.random.seed(42)
tf.random.set_seed(42)
# Need to have validation loss
early_stopping = keras.callbacks.EarlyStopping(monitor='val_loss',
min_delta=0.0,
patience=100,
verbose=2,
restore_best_weights=True)
encoder = keras.models.Sequential([keras.layers.Dense(200, input_shape=[800]),
keras.layers.Dense(15)])
decoder = keras.models.Sequential([keras.layers.Dense(200, input_shape=[15]),
keras.layers.Dense(800),
])
autoencoder = keras.models.Sequential([encoder, decoder])
autoencoder.compile(loss="mse",
optimizer=keras.optimizers.Nadam(lr=0.0003, beta_1=0.9, beta_2=0.999)
)
encoder.summary()
decoder.summary()
Model: "sequential_3" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= dense_2 (Dense) (None, 200) 160200 _________________________________________________________________ dense_3 (Dense) (None, 15) 3015 ================================================================= Total params: 163,215 Trainable params: 163,215 Non-trainable params: 0 _________________________________________________________________ Model: "sequential_4" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= dense_4 (Dense) (None, 200) 3200 _________________________________________________________________ dense_5 (Dense) (None, 800) 160800 ================================================================= Total params: 164,000 Trainable params: 164,000 Non-trainable params: 0 _________________________________________________________________
In [38]:
history = autoencoder.fit(X_train_1D_norm,
X_train_1D_norm,
epochs=1000,
validation_data=(X_train_1D_norm, X_train_1D_norm),
callbacks=[early_stopping])
Train on 3888 samples, validate on 3888 samples Epoch 1/1000 3888/3888 [==============================] - 1s 335us/sample - loss: 0.0535 - val_loss: 0.0152 Epoch 2/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 0.0083 - val_loss: 0.0040 Epoch 3/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 0.0029 - val_loss: 0.0019 Epoch 4/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 0.0019 - val_loss: 0.0017 Epoch 5/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 0.0018 - val_loss: 0.0019 Epoch 6/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 0.0016 - val_loss: 0.0012 Epoch 7/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 0.0013 - val_loss: 7.5517e-04 Epoch 8/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.8562e-04 - val_loss: 4.4975e-04 Epoch 9/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.9152e-04 - val_loss: 3.6823e-04 Epoch 10/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.6893e-04 - val_loss: 1.9823e-04 Epoch 11/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.7908e-04 - val_loss: 1.0087e-04 Epoch 12/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.6507e-04 - val_loss: 7.1798e-05 Epoch 13/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0696e-04 - val_loss: 6.1843e-05 Epoch 14/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.4705e-04 - val_loss: 4.1281e-05 Epoch 15/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 6.1790e-05 - val_loss: 4.8645e-05 Epoch 16/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.1749e-04 - val_loss: 7.6316e-05 Epoch 17/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.7837e-05 - val_loss: 3.9668e-05 Epoch 18/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.6528e-05 - val_loss: 3.8870e-05 Epoch 19/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 3.4340e-04 - val_loss: 5.0968e-05 Epoch 20/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.3672e-04 - val_loss: 3.5644e-05 Epoch 21/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.9512e-05 - val_loss: 3.8801e-05 Epoch 22/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.8148e-05 - val_loss: 3.9136e-05 Epoch 23/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.3418e-04 - val_loss: 3.5500e-05 Epoch 24/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 4.4832e-05 - val_loss: 5.4451e-05 Epoch 25/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.3332e-04 - val_loss: 4.1293e-05 Epoch 26/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.3092e-04 - val_loss: 4.5489e-05 Epoch 27/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.1948e-05 - val_loss: 4.6868e-05 Epoch 28/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.2867e-04 - val_loss: 1.0597e-04 Epoch 29/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.4215e-05 - val_loss: 4.1156e-05 Epoch 30/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.2528e-04 - val_loss: 9.1212e-04 Epoch 31/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 7.5877e-05 - val_loss: 3.5619e-05 Epoch 32/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 7.0622e-05 - val_loss: 9.7798e-05 Epoch 33/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.2486e-04 - val_loss: 3.8138e-05 Epoch 34/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 4.1571e-05 - val_loss: 3.9389e-05 Epoch 35/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.4383e-04 - val_loss: 3.6089e-05 Epoch 36/1000 3888/3888 [==============================] - 1s 186us/sample - loss: 6.6203e-05 - val_loss: 3.4543e-05 Epoch 37/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 7.5243e-05 - val_loss: 3.5585e-05 Epoch 38/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.3659e-05 - val_loss: 3.7065e-05 Epoch 39/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.9899e-05 - val_loss: 8.9075e-05 Epoch 40/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.8070e-04 - val_loss: 3.7009e-05 Epoch 41/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 7.0889e-05 - val_loss: 4.4309e-05 Epoch 42/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.9950e-04 - val_loss: 0.0023 Epoch 43/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.7480e-04 - val_loss: 3.2982e-05 Epoch 44/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.3070e-05 - val_loss: 4.2519e-05 Epoch 45/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.6713e-05 - val_loss: 6.0048e-05 Epoch 46/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0660e-04 - val_loss: 3.1515e-05 Epoch 47/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.2796e-04 - val_loss: 3.8759e-05 Epoch 48/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.9431e-05 - val_loss: 5.1361e-05 Epoch 49/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 8.8270e-05 - val_loss: 3.0570e-05 Epoch 50/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 3.7611e-05 - val_loss: 5.8796e-05 Epoch 51/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.7246e-05 - val_loss: 3.0547e-05 Epoch 52/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 7.4366e-05 - val_loss: 3.2117e-05 Epoch 53/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.5940e-05 - val_loss: 5.6891e-05 Epoch 54/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 7.7723e-05 - val_loss: 4.0344e-05 Epoch 55/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.5169e-05 - val_loss: 2.2869e-05 Epoch 56/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.4925e-05 - val_loss: 2.0909e-04 Epoch 57/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.5720e-04 - val_loss: 2.5054e-05 Epoch 58/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 4.5872e-05 - val_loss: 0.0011 Epoch 59/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 5.3317e-05 - val_loss: 2.0139e-05 Epoch 60/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.2335e-05 - val_loss: 2.0200e-05 Epoch 61/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 4.2515e-05 - val_loss: 2.4973e-05 Epoch 62/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 8.6700e-05 - val_loss: 2.7531e-05 Epoch 63/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.9338e-05 - val_loss: 1.9466e-05 Epoch 64/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 4.6554e-05 - val_loss: 2.1961e-05 Epoch 65/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.4244e-05 - val_loss: 4.8934e-05 Epoch 66/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.2925e-05 - val_loss: 1.8760e-05 Epoch 67/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.3063e-04 - val_loss: 8.5756e-05 Epoch 68/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.7991e-05 - val_loss: 1.6724e-05 Epoch 69/1000 3888/3888 [==============================] - 1s 187us/sample - loss: 5.3857e-05 - val_loss: 2.0216e-05 Epoch 70/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 4.5070e-05 - val_loss: 1.5495e-05 Epoch 71/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.6068e-05 - val_loss: 1.5344e-05 Epoch 72/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.5351e-05 - val_loss: 1.9476e-05 Epoch 73/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 4.6672e-05 - val_loss: 1.6797e-05 Epoch 74/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.6412e-05 - val_loss: 1.5885e-05 Epoch 75/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.5536e-05 - val_loss: 1.8104e-05 Epoch 76/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.8506e-05 - val_loss: 1.6382e-05 Epoch 77/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 7.5506e-05 - val_loss: 1.4393e-05 Epoch 78/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.6173e-05 - val_loss: 1.3941e-05 Epoch 79/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.3019e-05 - val_loss: 0.0020 Epoch 80/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.5053e-04 - val_loss: 1.4039e-05 Epoch 81/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.7747e-05 - val_loss: 1.2500e-05 Epoch 82/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.1950e-05 - val_loss: 1.3004e-05 Epoch 83/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.2864e-05 - val_loss: 1.2150e-05 Epoch 84/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.2116e-05 - val_loss: 1.1974e-05 Epoch 85/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.9654e-05 - val_loss: 1.3497e-05 Epoch 86/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.9995e-05 - val_loss: 1.1237e-05 Epoch 87/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 3.0470e-05 - val_loss: 1.4938e-05 Epoch 88/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.1920e-05 - val_loss: 5.1851e-04 Epoch 89/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.8859e-05 - val_loss: 9.0726e-06 Epoch 90/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.0550e-05 - val_loss: 2.3195e-05 Epoch 91/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.5695e-05 - val_loss: 8.0660e-06 Epoch 92/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.1090e-05 - val_loss: 5.1020e-05 Epoch 93/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 6.0481e-05 - val_loss: 7.8439e-06 Epoch 94/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 8.8006e-06 - val_loss: 9.1139e-05 Epoch 95/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 9.9224e-05 - val_loss: 8.4015e-06 Epoch 96/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.7083e-06 - val_loss: 8.9406e-06 Epoch 97/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 8.9340e-06 - val_loss: 1.2880e-05 Epoch 98/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.9769e-05 - val_loss: 8.0103e-06 Epoch 99/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.2969e-05 - val_loss: 8.7050e-06 Epoch 100/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 8.7083e-06 - val_loss: 8.9494e-06 Epoch 101/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 1.7718e-05 - val_loss: 2.2135e-05 Epoch 102/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 3.7573e-05 - val_loss: 6.2436e-06 Epoch 103/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.1693e-06 - val_loss: 7.1479e-06 Epoch 104/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.6742e-05 - val_loss: 1.6195e-05 Epoch 105/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.7330e-05 - val_loss: 1.3771e-05 Epoch 106/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 6.6983e-05 - val_loss: 1.2962e-05 Epoch 107/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.2605e-06 - val_loss: 5.8326e-06 Epoch 108/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 9.1957e-06 - val_loss: 6.8343e-06 Epoch 109/1000 3888/3888 [==============================] - 1s 184us/sample - loss: 2.9360e-05 - val_loss: 5.9915e-06 Epoch 110/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 6.9871e-05 - val_loss: 3.8160e-05 Epoch 111/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.4912e-05 - val_loss: 6.0910e-06 Epoch 112/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.8179e-06 - val_loss: 6.3737e-06 Epoch 113/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.1265e-05 - val_loss: 3.2333e-05 Epoch 114/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.1676e-05 - val_loss: 1.0195e-05 Epoch 115/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.6807e-05 - val_loss: 5.6441e-06 Epoch 116/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 4.4121e-05 - val_loss: 2.3282e-04 Epoch 117/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.4914e-05 - val_loss: 5.2071e-06 Epoch 118/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.2556e-05 - val_loss: 5.1676e-06 Epoch 119/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.4886e-06 - val_loss: 8.5413e-06 Epoch 120/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.2591e-05 - val_loss: 6.2924e-06 Epoch 121/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 9.2851e-06 - val_loss: 2.4171e-04 Epoch 122/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.9253e-05 - val_loss: 5.4999e-06 Epoch 123/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 6.3465e-06 - val_loss: 5.3613e-06 Epoch 124/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.5078e-05 - val_loss: 5.4000e-06 Epoch 125/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.7069e-06 - val_loss: 5.3014e-06 Epoch 126/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.7828e-05 - val_loss: 5.2466e-06 Epoch 127/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 5.6004e-06 - val_loss: 5.9656e-05 Epoch 128/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.7135e-05 - val_loss: 4.9637e-06 Epoch 129/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 2.4982e-05 - val_loss: 2.3459e-05 Epoch 130/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 7.3666e-06 - val_loss: 6.2106e-06 Epoch 131/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 3.0896e-05 - val_loss: 5.5841e-06 Epoch 132/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0555e-05 - val_loss: 9.0032e-06 Epoch 133/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.4195e-05 - val_loss: 1.2829e-04 Epoch 134/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.6516e-05 - val_loss: 5.1498e-06 Epoch 135/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.2393e-06 - val_loss: 8.0525e-06 Epoch 136/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.5683e-05 - val_loss: 1.1304e-05 Epoch 137/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.1194e-05 - val_loss: 1.0511e-04 Epoch 138/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 7.0561e-05 - val_loss: 5.7109e-06 Epoch 139/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.4951e-06 - val_loss: 1.3218e-05 Epoch 140/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 9.8751e-06 - val_loss: 7.0158e-06 Epoch 141/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.3423e-05 - val_loss: 5.6475e-05 Epoch 142/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 9.1158e-06 - val_loss: 5.5077e-06 Epoch 143/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.2764e-05 - val_loss: 4.9976e-06 Epoch 144/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.2453e-05 - val_loss: 8.4360e-06 Epoch 145/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.0160e-05 - val_loss: 4.5491e-06 Epoch 146/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 7.5740e-06 - val_loss: 4.5006e-06 Epoch 147/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.7311e-05 - val_loss: 5.0322e-06 Epoch 148/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.3695e-05 - val_loss: 3.8241e-05 Epoch 149/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.1614e-05 - val_loss: 5.3102e-06 Epoch 150/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0683e-05 - val_loss: 9.5388e-06 Epoch 151/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.6025e-05 - val_loss: 1.0934e-04 Epoch 152/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.3098e-05 - val_loss: 7.0422e-06 Epoch 153/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0359e-05 - val_loss: 5.0111e-06 Epoch 154/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 9.8739e-06 - val_loss: 5.0948e-06 Epoch 155/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.4709e-05 - val_loss: 4.4007e-06 Epoch 156/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 7.8001e-06 - val_loss: 3.3455e-05 Epoch 157/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.7936e-05 - val_loss: 4.1190e-06 Epoch 158/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.4802e-06 - val_loss: 6.3437e-06 Epoch 159/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.3475e-05 - val_loss: 1.2931e-05 Epoch 160/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.6131e-05 - val_loss: 3.2522e-05 Epoch 161/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.7984e-05 - val_loss: 6.7982e-06 Epoch 162/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 8.5168e-06 - val_loss: 4.5795e-06 Epoch 163/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 7.4001e-06 - val_loss: 4.9758e-06 Epoch 164/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 4.2477e-05 - val_loss: 4.3723e-06 Epoch 165/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 4.6957e-06 - val_loss: 2.1116e-05 Epoch 166/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 3.3955e-05 - val_loss: 4.0128e-06 Epoch 167/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 4.1741e-06 - val_loss: 8.2640e-06 Epoch 168/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.3437e-05 - val_loss: 4.7008e-06 Epoch 169/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 5.2707e-06 - val_loss: 1.0479e-05 Epoch 170/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.1098e-05 - val_loss: 6.0238e-05 Epoch 171/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.5370e-05 - val_loss: 4.2272e-06 Epoch 172/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 2.5175e-05 - val_loss: 2.4935e-05 Epoch 173/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 9.5772e-06 - val_loss: 4.1929e-06 Epoch 174/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 7.8287e-06 - val_loss: 2.0541e-05 Epoch 175/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 3.3390e-05 - val_loss: 3.8244e-05 Epoch 176/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 5.3276e-06 - val_loss: 1.2407e-05 Epoch 177/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 3.6688e-05 - val_loss: 5.6262e-06 Epoch 178/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 5.1916e-06 - val_loss: 4.3469e-06 Epoch 179/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.9968e-06 - val_loss: 4.9098e-06 Epoch 180/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.4352e-05 - val_loss: 5.6497e-06 Epoch 181/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 4.2309e-06 - val_loss: 5.1506e-06 Epoch 182/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 7.8824e-06 - val_loss: 3.6288e-06 Epoch 183/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 7.3779e-06 - val_loss: 8.7545e-06 Epoch 184/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.9285e-05 - val_loss: 3.4431e-06 Epoch 185/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 3.9090e-06 - val_loss: 2.6903e-05 Epoch 186/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.0726e-05 - val_loss: 7.9060e-06 Epoch 187/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.3175e-05 - val_loss: 4.2922e-06 Epoch 188/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.8306e-05 - val_loss: 3.9385e-06 Epoch 189/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.1134e-05 - val_loss: 4.0826e-06 Epoch 190/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.9324e-06 - val_loss: 1.5944e-04 Epoch 191/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 3.0891e-05 - val_loss: 3.2172e-06 Epoch 192/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.9906e-06 - val_loss: 4.1166e-06 Epoch 193/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.3886e-05 - val_loss: 4.9027e-06 Epoch 194/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.8207e-05 - val_loss: 3.5072e-06 Epoch 195/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 4.4939e-06 - val_loss: 3.5884e-06 Epoch 196/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.5110e-05 - val_loss: 5.6683e-06 Epoch 197/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.6174e-05 - val_loss: 3.4820e-06 Epoch 198/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 6.7733e-06 - val_loss: 6.7507e-05 Epoch 199/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.8176e-05 - val_loss: 5.0390e-06 Epoch 200/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.9197e-06 - val_loss: 3.7241e-06 Epoch 201/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2783e-05 - val_loss: 1.7701e-05 Epoch 202/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.3542e-06 - val_loss: 3.5766e-06 Epoch 203/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 4.4166e-05 - val_loss: 4.8503e-06 Epoch 204/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 4.4666e-06 - val_loss: 3.3946e-06 Epoch 205/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 8.5782e-06 - val_loss: 3.8985e-06 Epoch 206/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.7951e-06 - val_loss: 5.4282e-06 Epoch 207/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 1.7579e-05 - val_loss: 7.0563e-06 Epoch 208/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 8.9295e-06 - val_loss: 4.7452e-06 Epoch 209/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 1.2439e-05 - val_loss: 3.6179e-06 Epoch 210/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 4.7511e-06 - val_loss: 5.4727e-06 Epoch 211/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 2.7974e-05 - val_loss: 2.6158e-05 Epoch 212/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.5980e-05 - val_loss: 3.7714e-06 Epoch 213/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 5.4338e-06 - val_loss: 3.5042e-06 Epoch 214/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.4549e-05 - val_loss: 3.9779e-06 Epoch 215/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 5.7172e-06 - val_loss: 6.5461e-06 Epoch 216/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 3.4467e-05 - val_loss: 3.2974e-06 Epoch 217/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 8.1501e-06 - val_loss: 3.6679e-06 Epoch 218/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.2534e-06 - val_loss: 3.3477e-06 Epoch 219/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 7.9117e-05 - val_loss: 3.4949e-05 Epoch 220/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.7009e-05 - val_loss: 5.0496e-06 Epoch 221/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.7575e-06 - val_loss: 3.3998e-06 Epoch 222/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.6536e-06 - val_loss: 4.7240e-06 Epoch 223/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.3516e-05 - val_loss: 3.8390e-06 Epoch 224/1000 3888/3888 [==============================] - 1s 186us/sample - loss: 3.4694e-06 - val_loss: 3.2072e-06 Epoch 225/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 1.0708e-05 - val_loss: 3.3626e-06 Epoch 226/1000 3888/3888 [==============================] - 1s 187us/sample - loss: 8.0721e-06 - val_loss: 3.9620e-06 Epoch 227/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 7.0289e-06 - val_loss: 3.2581e-06 Epoch 228/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.9292e-05 - val_loss: 7.3522e-06 Epoch 229/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 9.6425e-06 - val_loss: 1.3922e-05 Epoch 230/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.9744e-06 - val_loss: 4.8593e-06 Epoch 231/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.0969e-05 - val_loss: 1.8333e-05 Epoch 232/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.9244e-05 - val_loss: 3.0772e-06 Epoch 233/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.9642e-06 - val_loss: 5.8316e-06 Epoch 234/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.9073e-06 - val_loss: 4.0091e-06 Epoch 235/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.4149e-06 - val_loss: 5.1649e-06 Epoch 236/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.9171e-05 - val_loss: 4.0765e-05 Epoch 237/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.1365e-05 - val_loss: 2.9481e-06 Epoch 238/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 8.3525e-06 - val_loss: 6.1301e-05 Epoch 239/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.5545e-05 - val_loss: 4.8968e-05 Epoch 240/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.3042e-05 - val_loss: 2.7705e-06 Epoch 241/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.8420e-06 - val_loss: 2.1102e-05 Epoch 242/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 2.2589e-05 - val_loss: 1.0970e-04 Epoch 243/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.3135e-06 - val_loss: 3.6154e-06 Epoch 244/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.4153e-06 - val_loss: 2.9603e-06 Epoch 245/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.2659e-05 - val_loss: 3.3719e-06 Epoch 246/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.7925e-05 - val_loss: 5.1221e-06 Epoch 247/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.0594e-06 - val_loss: 2.8336e-06 Epoch 248/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.4678e-05 - val_loss: 4.5219e-06 Epoch 249/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 8.3124e-06 - val_loss: 6.9377e-06 Epoch 250/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.0095e-05 - val_loss: 3.0813e-06 Epoch 251/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 8.2342e-06 - val_loss: 3.0500e-06 Epoch 252/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.1783e-06 - val_loss: 2.7052e-06 Epoch 253/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 3.7919e-05 - val_loss: 3.1820e-06 Epoch 254/1000 3888/3888 [==============================] - 1s 191us/sample - loss: 3.0671e-06 - val_loss: 1.3160e-05 Epoch 255/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 7.3599e-06 - val_loss: 1.1802e-05 Epoch 256/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0280e-05 - val_loss: 7.9171e-05 Epoch 257/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.4619e-05 - val_loss: 3.1317e-06 Epoch 258/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0080e-05 - val_loss: 3.6632e-06 Epoch 259/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.5901e-06 - val_loss: 1.3614e-05 Epoch 260/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.3085e-05 - val_loss: 2.4088e-06 Epoch 261/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.5287e-06 - val_loss: 2.9948e-06 Epoch 262/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.7372e-05 - val_loss: 2.8046e-06 Epoch 263/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 5.6635e-06 - val_loss: 2.7423e-06 Epoch 264/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.7643e-05 - val_loss: 5.3140e-06 Epoch 265/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.1403e-05 - val_loss: 2.6308e-06 Epoch 266/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.1165e-05 - val_loss: 2.8155e-06 Epoch 267/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0328e-05 - val_loss: 6.9861e-05 Epoch 268/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.0866e-05 - val_loss: 4.2737e-06 Epoch 269/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.6277e-06 - val_loss: 4.7183e-06 Epoch 270/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.5881e-06 - val_loss: 5.0018e-06 Epoch 271/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 2.5373e-05 - val_loss: 2.4953e-06 Epoch 272/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.5949e-06 - val_loss: 6.7575e-06 Epoch 273/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 5.4496e-06 - val_loss: 9.0777e-05 Epoch 274/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0716e-05 - val_loss: 7.0088e-06 Epoch 275/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 6.1252e-06 - val_loss: 7.9399e-06 Epoch 276/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2819e-05 - val_loss: 1.1203e-05 Epoch 277/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.4371e-05 - val_loss: 4.4543e-06 Epoch 278/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.2190e-05 - val_loss: 2.8272e-06 Epoch 279/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.1505e-06 - val_loss: 4.8731e-06 Epoch 280/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.8524e-05 - val_loss: 4.0713e-06 Epoch 281/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.4387e-06 - val_loss: 2.7308e-06 Epoch 282/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 4.2541e-06 - val_loss: 2.6044e-06 Epoch 283/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.1921e-05 - val_loss: 7.5142e-04 Epoch 284/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.0587e-05 - val_loss: 2.9926e-06 Epoch 285/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.5475e-06 - val_loss: 2.3648e-06 Epoch 286/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.5947e-06 - val_loss: 4.1283e-06 Epoch 287/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.9303e-06 - val_loss: 2.4575e-06 Epoch 288/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.8683e-05 - val_loss: 3.0071e-06 Epoch 289/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.4518e-06 - val_loss: 8.3101e-06 Epoch 290/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 5.8345e-06 - val_loss: 4.1932e-06 Epoch 291/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.4498e-05 - val_loss: 3.2619e-06 Epoch 292/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 6.6530e-06 - val_loss: 1.9060e-05 Epoch 293/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0283e-05 - val_loss: 3.6969e-04 Epoch 294/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.8000e-05 - val_loss: 7.3509e-06 Epoch 295/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.1389e-06 - val_loss: 3.0402e-06 Epoch 296/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.8336e-06 - val_loss: 1.7918e-05 Epoch 297/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.1507e-05 - val_loss: 1.3124e-05 Epoch 298/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.7472e-06 - val_loss: 1.4062e-05 Epoch 299/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.0254e-05 - val_loss: 3.7220e-06 Epoch 300/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.4374e-06 - val_loss: 2.3589e-06 Epoch 301/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 5.0143e-06 - val_loss: 2.9043e-05 Epoch 302/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.7488e-05 - val_loss: 2.6786e-06 Epoch 303/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.5928e-06 - val_loss: 3.9229e-06 Epoch 304/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.0839e-05 - val_loss: 2.4523e-06 Epoch 305/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.1102e-06 - val_loss: 2.1836e-06 Epoch 306/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 5.9010e-06 - val_loss: 8.9886e-06 Epoch 307/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 8.4445e-06 - val_loss: 2.3263e-06 Epoch 308/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.6847e-05 - val_loss: 2.3667e-06 Epoch 309/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0421e-05 - val_loss: 8.4373e-05 Epoch 310/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.3871e-06 - val_loss: 3.5441e-06 Epoch 311/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.0376e-05 - val_loss: 1.5477e-05 Epoch 312/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 3.2257e-06 - val_loss: 2.0410e-06 Epoch 313/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 7.8253e-06 - val_loss: 2.9242e-06 Epoch 314/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 2.3067e-06 - val_loss: 2.1962e-06 Epoch 315/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.8522e-05 - val_loss: 2.0109e-05 Epoch 316/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.7619e-06 - val_loss: 2.1300e-06 Epoch 317/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 6.8799e-06 - val_loss: 7.6182e-06 Epoch 318/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 8.5783e-06 - val_loss: 2.3599e-06 Epoch 319/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 7.3191e-06 - val_loss: 6.5161e-06 Epoch 320/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 7.6015e-06 - val_loss: 2.1173e-06 Epoch 321/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 8.1310e-06 - val_loss: 2.0808e-05 Epoch 322/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 9.7476e-06 - val_loss: 1.4807e-05 Epoch 323/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.3967e-05 - val_loss: 2.0725e-06 Epoch 324/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.6539e-06 - val_loss: 2.3632e-06 Epoch 325/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.4266e-05 - val_loss: 3.0255e-06 Epoch 326/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.6503e-06 - val_loss: 2.3852e-06 Epoch 327/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.8080e-05 - val_loss: 3.0900e-06 Epoch 328/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.4382e-06 - val_loss: 2.9060e-06 Epoch 329/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 9.3559e-06 - val_loss: 5.1858e-06 Epoch 330/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 9.0298e-06 - val_loss: 2.3302e-06 Epoch 331/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 2.3178e-06 - val_loss: 2.2415e-06 Epoch 332/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.8374e-05 - val_loss: 5.9787e-06 Epoch 333/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 4.1784e-06 - val_loss: 1.7912e-06 Epoch 334/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 6.7554e-06 - val_loss: 2.2403e-06 Epoch 335/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0839e-05 - val_loss: 1.0753e-04 Epoch 336/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.9268e-05 - val_loss: 2.3432e-06 Epoch 337/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 2.7413e-06 - val_loss: 2.7888e-06 Epoch 338/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.1672e-05 - val_loss: 5.6468e-06 Epoch 339/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.1170e-06 - val_loss: 1.8866e-06 Epoch 340/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.8403e-06 - val_loss: 9.0113e-06 Epoch 341/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.8078e-05 - val_loss: 2.2221e-06 Epoch 342/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.0600e-06 - val_loss: 2.1397e-05 Epoch 343/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.5691e-05 - val_loss: 1.0458e-05 Epoch 344/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.1454e-06 - val_loss: 1.8779e-06 Epoch 345/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.1160e-05 - val_loss: 5.8678e-06 Epoch 346/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.0822e-06 - val_loss: 2.4306e-06 Epoch 347/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 6.6364e-06 - val_loss: 3.7464e-06 Epoch 348/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 9.7320e-06 - val_loss: 3.6804e-06 Epoch 349/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 4.5874e-06 - val_loss: 2.0960e-06 Epoch 350/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 2.2424e-05 - val_loss: 4.8930e-06 Epoch 351/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 6.4397e-06 - val_loss: 1.9092e-06 Epoch 352/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 2.5988e-06 - val_loss: 3.8970e-06 Epoch 353/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 1.3001e-05 - val_loss: 1.1140e-05 Epoch 354/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 9.9668e-06 - val_loss: 1.7087e-06 Epoch 355/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.2530e-06 - val_loss: 2.1067e-06 Epoch 356/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.3583e-05 - val_loss: 1.5536e-05 Epoch 357/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.3133e-06 - val_loss: 1.9105e-06 Epoch 358/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.2270e-05 - val_loss: 8.5298e-06 Epoch 359/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 8.4059e-06 - val_loss: 6.9142e-06 Epoch 360/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.0029e-06 - val_loss: 2.9535e-06 Epoch 361/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.3502e-06 - val_loss: 2.2446e-04 Epoch 362/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 7.7992e-06 - val_loss: 1.0232e-05 Epoch 363/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.2512e-05 - val_loss: 2.1387e-06 Epoch 364/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.5206e-05 - val_loss: 2.6470e-06 Epoch 365/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.8718e-06 - val_loss: 1.7047e-06 Epoch 366/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.2842e-06 - val_loss: 2.4473e-06 Epoch 367/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 7.6799e-06 - val_loss: 4.1565e-05 Epoch 368/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 9.8564e-06 - val_loss: 9.8573e-06 Epoch 369/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.9651e-06 - val_loss: 7.1640e-06 Epoch 370/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.0943e-05 - val_loss: 2.0081e-05 Epoch 371/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.8441e-06 - val_loss: 1.5971e-06 Epoch 372/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.8798e-06 - val_loss: 2.0005e-06 Epoch 373/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.5734e-06 - val_loss: 1.6462e-06 Epoch 374/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.8593e-06 - val_loss: 1.2738e-05 Epoch 375/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.4957e-05 - val_loss: 1.8044e-06 Epoch 376/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.5031e-06 - val_loss: 3.2377e-06 Epoch 377/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.9384e-06 - val_loss: 1.7004e-06 Epoch 378/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.4203e-05 - val_loss: 1.4514e-05 Epoch 379/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.1299e-06 - val_loss: 4.3339e-06 Epoch 380/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.2026e-06 - val_loss: 2.5315e-06 Epoch 381/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.3867e-05 - val_loss: 3.9240e-06 Epoch 382/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.0593e-06 - val_loss: 1.8143e-06 Epoch 383/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.7075e-06 - val_loss: 1.7494e-06 Epoch 384/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.6606e-06 - val_loss: 1.6647e-06 Epoch 385/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 6.9009e-06 - val_loss: 2.5021e-06 Epoch 386/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 4.2177e-06 - val_loss: 2.3959e-06 Epoch 387/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.0853e-06 - val_loss: 7.7081e-06 Epoch 388/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 1.8919e-05 - val_loss: 2.4233e-06 Epoch 389/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 2.9971e-06 - val_loss: 1.9259e-06 Epoch 390/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.9731e-06 - val_loss: 1.5496e-06 Epoch 391/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 3.7260e-06 - val_loss: 1.4467e-06 Epoch 392/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.0513e-05 - val_loss: 3.8704e-06 Epoch 393/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.3180e-06 - val_loss: 1.5908e-06 Epoch 394/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.2209e-06 - val_loss: 1.6939e-06 Epoch 395/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.4494e-05 - val_loss: 2.9699e-06 Epoch 396/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.0743e-05 - val_loss: 5.9139e-05 Epoch 397/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 3.6338e-05 - val_loss: 2.1397e-06 Epoch 398/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.7930e-06 - val_loss: 1.8215e-06 Epoch 399/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.7690e-06 - val_loss: 1.8423e-06 Epoch 400/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.7020e-06 - val_loss: 1.5541e-06 Epoch 401/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.6110e-06 - val_loss: 2.3584e-06 Epoch 402/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.8290e-05 - val_loss: 1.5521e-06 Epoch 403/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.5587e-06 - val_loss: 2.4611e-06 Epoch 404/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.1388e-06 - val_loss: 2.5938e-06 Epoch 405/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 9.9370e-06 - val_loss: 1.4917e-06 Epoch 406/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 4.0052e-06 - val_loss: 2.6830e-06 Epoch 407/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 5.3743e-06 - val_loss: 2.2690e-05 Epoch 408/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.5592e-05 - val_loss: 1.5562e-06 Epoch 409/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.5428e-06 - val_loss: 2.7522e-06 Epoch 410/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.7896e-06 - val_loss: 3.0423e-06 Epoch 411/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 2.1510e-06 - val_loss: 1.5495e-05 Epoch 412/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 6.0791e-06 - val_loss: 1.4799e-06 Epoch 413/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.0433e-05 - val_loss: 1.3086e-06 Epoch 414/1000 3888/3888 [==============================] - 1s 190us/sample - loss: 1.7596e-06 - val_loss: 4.1308e-06 Epoch 415/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.5397e-05 - val_loss: 4.2626e-05 Epoch 416/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.6931e-06 - val_loss: 1.5022e-06 Epoch 417/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 3.5351e-05 - val_loss: 3.3211e-06 Epoch 418/1000 3888/3888 [==============================] - 1s 185us/sample - loss: 1.4503e-06 - val_loss: 1.1680e-06 Epoch 419/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.6460e-06 - val_loss: 1.2937e-06 Epoch 420/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.6762e-06 - val_loss: 1.3255e-06 Epoch 421/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.1399e-05 - val_loss: 1.5501e-06 Epoch 422/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.0778e-06 - val_loss: 1.1850e-06 Epoch 423/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0573e-05 - val_loss: 1.7515e-05 Epoch 424/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.0132e-06 - val_loss: 1.3778e-06 Epoch 425/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.7217e-06 - val_loss: 1.0759e-05 Epoch 426/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.7324e-06 - val_loss: 7.7046e-06 Epoch 427/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.5185e-05 - val_loss: 1.9901e-06 Epoch 428/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0865e-05 - val_loss: 1.3619e-05 Epoch 429/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.2381e-06 - val_loss: 3.9273e-06 Epoch 430/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 9.2677e-06 - val_loss: 1.2386e-06 Epoch 431/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0268e-05 - val_loss: 1.0382e-05 Epoch 432/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.2590e-06 - val_loss: 1.7713e-06 Epoch 433/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.6035e-06 - val_loss: 2.0373e-06 Epoch 434/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.3752e-06 - val_loss: 2.4552e-06 Epoch 435/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.7909e-05 - val_loss: 7.3176e-05 Epoch 436/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 6.5958e-06 - val_loss: 1.9673e-06 Epoch 437/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.8889e-06 - val_loss: 1.5430e-06 Epoch 438/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.2388e-06 - val_loss: 9.4830e-06 Epoch 439/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.4697e-06 - val_loss: 4.9742e-06 Epoch 440/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.2152e-05 - val_loss: 1.7234e-06 Epoch 441/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.3063e-06 - val_loss: 2.4266e-06 Epoch 442/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.8344e-06 - val_loss: 1.4443e-06 Epoch 443/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.8414e-05 - val_loss: 1.9922e-06 Epoch 444/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 8.0730e-06 - val_loss: 1.4204e-06 Epoch 445/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.1498e-05 - val_loss: 4.2393e-05 Epoch 446/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.0319e-06 - val_loss: 1.8748e-06 Epoch 447/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.1341e-06 - val_loss: 4.3277e-06 Epoch 448/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.0909e-05 - val_loss: 2.3485e-06 Epoch 449/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2065e-06 - val_loss: 1.1550e-06 Epoch 450/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.5939e-06 - val_loss: 6.7722e-06 Epoch 451/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0346e-05 - val_loss: 1.2055e-06 Epoch 452/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0227e-05 - val_loss: 1.3773e-05 Epoch 453/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 3.0813e-06 - val_loss: 1.2987e-06 Epoch 454/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.5758e-06 - val_loss: 2.7455e-04 Epoch 455/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.5428e-05 - val_loss: 1.4038e-06 Epoch 456/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.3902e-06 - val_loss: 1.8620e-06 Epoch 457/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.3376e-06 - val_loss: 1.1622e-05 Epoch 458/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.0708e-05 - val_loss: 1.5163e-06 Epoch 459/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.6082e-06 - val_loss: 1.6680e-06 Epoch 460/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.5369e-06 - val_loss: 1.2157e-06 Epoch 461/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 5.6462e-06 - val_loss: 1.4192e-06 Epoch 462/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.1915e-05 - val_loss: 1.0203e-06 Epoch 463/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.8084e-05 - val_loss: 1.0020e-06 Epoch 464/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.2712e-06 - val_loss: 3.6884e-06 Epoch 465/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 7.1238e-06 - val_loss: 3.9790e-06 Epoch 466/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.2595e-05 - val_loss: 5.2550e-06 Epoch 467/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 7.1389e-06 - val_loss: 9.7097e-06 Epoch 468/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.4407e-06 - val_loss: 9.6495e-07 Epoch 469/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.4700e-06 - val_loss: 1.8017e-06 Epoch 470/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.2227e-05 - val_loss: 1.1400e-06 Epoch 471/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 2.9678e-06 - val_loss: 9.4501e-06 Epoch 472/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.1334e-05 - val_loss: 1.1184e-06 Epoch 473/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 2.5644e-06 - val_loss: 1.3535e-06 Epoch 474/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 2.0811e-05 - val_loss: 2.7761e-06 Epoch 475/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.0822e-06 - val_loss: 9.1740e-07 Epoch 476/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.3874e-06 - val_loss: 1.0856e-06 Epoch 477/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 6.1273e-06 - val_loss: 3.6102e-06 Epoch 478/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 5.2369e-06 - val_loss: 1.5516e-06 Epoch 479/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 1.2305e-05 - val_loss: 1.0515e-06 Epoch 480/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.4324e-06 - val_loss: 1.0629e-06 Epoch 481/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 2.7596e-05 - val_loss: 4.1984e-06 Epoch 482/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.2255e-06 - val_loss: 1.2940e-06 Epoch 483/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.1352e-06 - val_loss: 1.4913e-06 Epoch 484/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.5321e-06 - val_loss: 1.4226e-06 Epoch 485/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.1353e-05 - val_loss: 9.2207e-07 Epoch 486/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.1660e-06 - val_loss: 1.5253e-06 Epoch 487/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 1.8730e-05 - val_loss: 7.4030e-06 Epoch 488/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 1.7292e-06 - val_loss: 1.2631e-06 Epoch 489/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 2.5786e-06 - val_loss: 1.0444e-06 Epoch 490/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 2.8469e-06 - val_loss: 4.4975e-06 Epoch 491/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 2.7076e-05 - val_loss: 1.6874e-06 Epoch 492/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 1.0502e-06 - val_loss: 1.0931e-06 Epoch 493/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 1.2618e-05 - val_loss: 1.8742e-06 Epoch 494/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 2.8099e-06 - val_loss: 9.8976e-07 Epoch 495/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.9711e-06 - val_loss: 9.0870e-07 Epoch 496/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 3.3500e-06 - val_loss: 1.3082e-06 Epoch 497/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.0904e-05 - val_loss: 1.5501e-05 Epoch 498/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.2045e-05 - val_loss: 4.6749e-06 Epoch 499/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 5.6491e-06 - val_loss: 9.2892e-07 Epoch 500/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.0366e-06 - val_loss: 1.0301e-06 Epoch 501/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.2371e-05 - val_loss: 1.1822e-06 Epoch 502/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 1.8258e-06 - val_loss: 1.0629e-06 Epoch 503/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 4.6098e-06 - val_loss: 3.5243e-05 Epoch 504/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 8.5696e-06 - val_loss: 8.1164e-06 Epoch 505/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 7.9431e-06 - val_loss: 1.1133e-06 Epoch 506/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.8857e-06 - val_loss: 7.0038e-06 Epoch 507/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 2.1959e-05 - val_loss: 9.0636e-07 Epoch 508/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 1.1916e-06 - val_loss: 8.9921e-07 Epoch 509/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 9.4945e-06 - val_loss: 2.0001e-04 Epoch 510/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 6.8017e-06 - val_loss: 1.1831e-06 Epoch 511/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.1870e-05 - val_loss: 1.0599e-06 Epoch 512/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.7783e-06 - val_loss: 2.8202e-06 Epoch 513/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 4.4915e-06 - val_loss: 8.2696e-06 Epoch 514/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.6106e-05 - val_loss: 2.2567e-06 Epoch 515/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.1247e-06 - val_loss: 1.2957e-06 Epoch 516/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.3345e-06 - val_loss: 7.6377e-06 Epoch 517/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.4168e-05 - val_loss: 1.5902e-06 Epoch 518/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.7137e-06 - val_loss: 1.1840e-06 Epoch 519/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.6438e-05 - val_loss: 1.8122e-06 Epoch 520/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 2.4871e-06 - val_loss: 1.4797e-06 Epoch 521/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 4.5490e-06 - val_loss: 8.0070e-07 Epoch 522/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 3.4803e-05 - val_loss: 1.6581e-06 Epoch 523/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 1.1841e-06 - val_loss: 9.4026e-07 Epoch 524/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.0951e-06 - val_loss: 9.5357e-07 Epoch 525/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 3.1694e-06 - val_loss: 1.3367e-06 Epoch 526/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 5.8788e-06 - val_loss: 1.9314e-06 Epoch 527/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 5.3872e-06 - val_loss: 3.8112e-06 Epoch 528/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 5.6085e-06 - val_loss: 2.8063e-06 Epoch 529/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 2.0449e-06 - val_loss: 2.1868e-06 Epoch 530/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.1434e-05 - val_loss: 5.5874e-05 Epoch 531/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.4473e-05 - val_loss: 6.3840e-04 Epoch 532/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.1531e-05 - val_loss: 9.8177e-07 Epoch 533/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.2701e-06 - val_loss: 1.1586e-06 Epoch 534/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 3.9397e-06 - val_loss: 9.8483e-06 Epoch 535/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 4.9912e-06 - val_loss: 8.5707e-07 Epoch 536/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.5146e-06 - val_loss: 5.3427e-06 Epoch 537/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0996e-05 - val_loss: 3.2856e-06 Epoch 538/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.8692e-06 - val_loss: 1.7211e-05 Epoch 539/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.3567e-06 - val_loss: 9.3089e-07 Epoch 540/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.4488e-06 - val_loss: 1.7783e-06 Epoch 541/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.3697e-06 - val_loss: 4.2840e-05 Epoch 542/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.6074e-05 - val_loss: 9.5374e-07 Epoch 543/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.2452e-06 - val_loss: 5.2246e-06 Epoch 544/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.5425e-05 - val_loss: 2.3539e-06 Epoch 545/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.4071e-06 - val_loss: 4.8955e-06 Epoch 546/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 4.3799e-06 - val_loss: 1.3718e-05 Epoch 547/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.1709e-05 - val_loss: 6.7364e-06 Epoch 548/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.7424e-06 - val_loss: 9.7109e-07 Epoch 549/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.4036e-05 - val_loss: 6.8480e-06 Epoch 550/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.2266e-06 - val_loss: 1.0541e-06 Epoch 551/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.0686e-06 - val_loss: 2.0343e-06 Epoch 552/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 6.6201e-06 - val_loss: 1.8123e-06 Epoch 553/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.4260e-05 - val_loss: 8.2746e-06 Epoch 554/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.2413e-06 - val_loss: 1.7275e-06 Epoch 555/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.1350e-06 - val_loss: 2.8627e-06 Epoch 556/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.1295e-06 - val_loss: 1.0613e-06 Epoch 557/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.1084e-06 - val_loss: 1.9802e-06 Epoch 558/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 9.0013e-06 - val_loss: 3.9983e-05 Epoch 559/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.1063e-05 - val_loss: 2.9720e-06 Epoch 560/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.2798e-06 - val_loss: 1.1325e-06 Epoch 561/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.1539e-06 - val_loss: 6.9933e-07 Epoch 562/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.8772e-06 - val_loss: 1.5867e-06 Epoch 563/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.4767e-05 - val_loss: 1.0609e-06 Epoch 564/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.1623e-06 - val_loss: 7.6949e-07 Epoch 565/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.5857e-06 - val_loss: 4.4581e-06 Epoch 566/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.4753e-06 - val_loss: 7.1253e-07 Epoch 567/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.6149e-06 - val_loss: 1.3393e-06 Epoch 568/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.0726e-06 - val_loss: 8.3753e-06 Epoch 569/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.5192e-05 - val_loss: 5.1985e-06 Epoch 570/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.2698e-06 - val_loss: 1.0956e-06 Epoch 571/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 9.7896e-07 - val_loss: 1.3943e-06 Epoch 572/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.2841e-05 - val_loss: 1.0330e-06 Epoch 573/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.6088e-06 - val_loss: 1.7922e-05 Epoch 574/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.1366e-05 - val_loss: 1.1295e-06 Epoch 575/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 9.3262e-07 - val_loss: 8.2604e-07 Epoch 576/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.1485e-06 - val_loss: 1.9711e-06 Epoch 577/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 5.3565e-06 - val_loss: 1.1141e-06 Epoch 578/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.1675e-06 - val_loss: 1.1055e-06 Epoch 579/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0166e-05 - val_loss: 1.0169e-06 Epoch 580/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.4294e-06 - val_loss: 1.0586e-06 Epoch 581/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.3334e-06 - val_loss: 2.0306e-06 Epoch 582/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 9.7737e-06 - val_loss: 9.6329e-07 Epoch 583/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.3132e-06 - val_loss: 2.9515e-06 Epoch 584/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 8.3841e-06 - val_loss: 8.0121e-07 Epoch 585/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.3057e-06 - val_loss: 1.4063e-06 Epoch 586/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.3206e-05 - val_loss: 8.1323e-07 Epoch 587/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.0071e-06 - val_loss: 1.7559e-05 Epoch 588/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 6.4735e-06 - val_loss: 1.1089e-05 Epoch 589/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.0442e-05 - val_loss: 8.7763e-07 Epoch 590/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.7262e-06 - val_loss: 7.2418e-06 Epoch 591/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 9.4999e-06 - val_loss: 9.1392e-07 Epoch 592/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.9401e-06 - val_loss: 1.4601e-05 Epoch 593/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 6.0707e-06 - val_loss: 2.8764e-05 Epoch 594/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.2990e-05 - val_loss: 6.7774e-06 Epoch 595/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.4543e-05 - val_loss: 6.4575e-05 Epoch 596/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 2.1851e-06 - val_loss: 1.0611e-06 Epoch 597/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 4.2005e-06 - val_loss: 1.0698e-06 Epoch 598/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 3.7407e-06 - val_loss: 4.2647e-06 Epoch 599/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 2.2542e-06 - val_loss: 1.9349e-06 Epoch 600/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.3838e-05 - val_loss: 1.1409e-06 Epoch 601/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.5660e-06 - val_loss: 1.2980e-06 Epoch 602/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 9.9212e-06 - val_loss: 1.5549e-06 Epoch 603/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.0960e-06 - val_loss: 9.9821e-07 Epoch 604/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 2.5465e-05 - val_loss: 1.3675e-06 Epoch 605/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 8.3191e-07 - val_loss: 9.2478e-07 Epoch 606/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.4061e-06 - val_loss: 8.4134e-07 Epoch 607/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 1.9697e-06 - val_loss: 1.7388e-06 Epoch 608/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.7452e-05 - val_loss: 3.1190e-06 Epoch 609/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.4322e-06 - val_loss: 9.2444e-07 Epoch 610/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.5410e-06 - val_loss: 1.4310e-06 Epoch 611/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 6.1103e-06 - val_loss: 1.5369e-06 Epoch 612/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.5616e-06 - val_loss: 2.2030e-06 Epoch 613/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.1420e-05 - val_loss: 4.3294e-04 Epoch 614/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 9.9325e-06 - val_loss: 8.3826e-07 Epoch 615/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.4105e-06 - val_loss: 9.8381e-07 Epoch 616/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 9.9980e-06 - val_loss: 3.0689e-05 Epoch 617/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 5.2282e-06 - val_loss: 8.7572e-07 Epoch 618/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.9242e-06 - val_loss: 6.9489e-07 Epoch 619/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 4.9974e-06 - val_loss: 1.3678e-06 Epoch 620/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 9.2496e-06 - val_loss: 7.3139e-07 Epoch 621/1000 3888/3888 [==============================] - 1s 184us/sample - loss: 8.3775e-07 - val_loss: 1.2029e-06 Epoch 622/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 8.3227e-06 - val_loss: 2.7065e-05 Epoch 623/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 1.6273e-05 - val_loss: 6.7766e-07 Epoch 624/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 8.2391e-07 - val_loss: 7.2137e-07 Epoch 625/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 9.6558e-07 - val_loss: 7.7050e-07 Epoch 626/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.3097e-05 - val_loss: 1.5587e-06 Epoch 627/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.4848e-06 - val_loss: 8.8139e-07 Epoch 628/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.6687e-05 - val_loss: 1.7469e-06 Epoch 629/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 8.4280e-07 - val_loss: 7.7586e-07 Epoch 630/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 2.1256e-06 - val_loss: 2.8311e-06 Epoch 631/1000 3888/3888 [==============================] - 1s 184us/sample - loss: 1.8128e-05 - val_loss: 8.6366e-07 Epoch 632/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.0335e-06 - val_loss: 7.3205e-07 Epoch 633/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 4.5946e-06 - val_loss: 2.7754e-06 Epoch 634/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 3.6458e-06 - val_loss: 7.5278e-07 Epoch 635/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 5.1918e-06 - val_loss: 7.2803e-07 Epoch 636/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.2323e-06 - val_loss: 6.4537e-07 Epoch 637/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.7709e-06 - val_loss: 7.2236e-07 Epoch 638/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 1.5383e-05 - val_loss: 1.0185e-06 Epoch 639/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 7.7825e-06 - val_loss: 7.6979e-05 Epoch 640/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 6.3464e-06 - val_loss: 3.1776e-05 Epoch 641/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 8.1383e-06 - val_loss: 7.3504e-07 Epoch 642/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 7.0134e-07 - val_loss: 1.2657e-06 Epoch 643/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.8160e-06 - val_loss: 6.4372e-07 Epoch 644/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.8117e-05 - val_loss: 8.9588e-07 Epoch 645/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 8.2953e-07 - val_loss: 7.4985e-07 Epoch 646/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 1.2635e-06 - val_loss: 1.2836e-06 Epoch 647/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.0888e-05 - val_loss: 1.7435e-05 Epoch 648/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 4.8301e-06 - val_loss: 1.0895e-06 Epoch 649/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 9.9777e-07 - val_loss: 9.9591e-07 Epoch 650/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.9786e-07 - val_loss: 8.0778e-07 Epoch 651/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.3470e-06 - val_loss: 1.0469e-05 Epoch 652/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.0853e-06 - val_loss: 9.0029e-07 Epoch 653/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.6292e-06 - val_loss: 7.4468e-07 Epoch 654/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.5883e-05 - val_loss: 7.1694e-07 Epoch 655/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 9.6981e-07 - val_loss: 2.3582e-06 Epoch 656/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 9.5429e-07 - val_loss: 7.0387e-07 Epoch 657/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 8.6699e-06 - val_loss: 7.4278e-06 Epoch 658/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 3.7977e-06 - val_loss: 1.0903e-06 Epoch 659/1000 3888/3888 [==============================] - 1s 188us/sample - loss: 1.5148e-06 - val_loss: 3.1882e-06 Epoch 660/1000 3888/3888 [==============================] - 1s 183us/sample - loss: 1.3912e-05 - val_loss: 1.0949e-06 Epoch 661/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.5234e-06 - val_loss: 3.9370e-06 Epoch 662/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.0094e-05 - val_loss: 1.4405e-05 Epoch 663/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 2.0956e-06 - val_loss: 1.2220e-06 Epoch 664/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 1.5128e-06 - val_loss: 1.9773e-05 Epoch 665/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 6.9733e-06 - val_loss: 8.7898e-07 Epoch 666/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 8.7990e-06 - val_loss: 2.6586e-05 Epoch 667/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.1423e-06 - val_loss: 8.4490e-07 Epoch 668/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.1550e-05 - val_loss: 3.1634e-04 Epoch 669/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.7539e-06 - val_loss: 6.4227e-07 Epoch 670/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.8929e-07 - val_loss: 7.7161e-07 Epoch 671/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.8981e-06 - val_loss: 3.9827e-05 Epoch 672/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 6.9435e-06 - val_loss: 7.5519e-07 Epoch 673/1000 3888/3888 [==============================] - 1s 180us/sample - loss: 4.9098e-06 - val_loss: 7.8272e-07 Epoch 674/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.5963e-05 - val_loss: 1.1199e-05 Epoch 675/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.3209e-06 - val_loss: 7.1964e-07 Epoch 676/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.8819e-07 - val_loss: 7.7390e-07 Epoch 677/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 4.8237e-06 - val_loss: 9.5430e-07 Epoch 678/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.6823e-06 - val_loss: 1.2342e-06 Epoch 679/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.6249e-05 - val_loss: 7.6970e-07 Epoch 680/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.4951e-06 - val_loss: 1.7948e-06 Epoch 681/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.8288e-06 - val_loss: 1.9353e-06 Epoch 682/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.2006e-05 - val_loss: 6.8243e-07 Epoch 683/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.5268e-06 - val_loss: 2.6148e-05 Epoch 684/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 6.6634e-06 - val_loss: 6.8921e-07 Epoch 685/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.3126e-06 - val_loss: 1.3975e-06 Epoch 686/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0464e-06 - val_loss: 7.2380e-07 Epoch 687/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.2062e-05 - val_loss: 9.0180e-07 Epoch 688/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 7.5753e-07 - val_loss: 7.5016e-07 Epoch 689/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.1400e-06 - val_loss: 8.2846e-07 Epoch 690/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.2835e-06 - val_loss: 1.3102e-05 Epoch 691/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.4549e-06 - val_loss: 9.5385e-07 Epoch 692/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 5.2391e-06 - val_loss: 4.5766e-06 Epoch 693/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.6873e-06 - val_loss: 8.5413e-07 Epoch 694/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 8.3717e-06 - val_loss: 2.9855e-06 Epoch 695/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.1306e-06 - val_loss: 7.3996e-06 Epoch 696/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.3075e-06 - val_loss: 5.7341e-05 Epoch 697/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.8246e-06 - val_loss: 5.8163e-07 Epoch 698/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 6.7242e-06 - val_loss: 7.4032e-07 Epoch 699/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.4336e-06 - val_loss: 1.0061e-06 Epoch 700/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2117e-05 - val_loss: 2.3619e-06 Epoch 701/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.4556e-06 - val_loss: 7.4717e-07 Epoch 702/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.1042e-06 - val_loss: 1.8059e-05 Epoch 703/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.1900e-05 - val_loss: 1.5014e-06 Epoch 704/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 7.6848e-07 - val_loss: 7.6811e-07 Epoch 705/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.3493e-06 - val_loss: 9.0395e-07 Epoch 706/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.4063e-06 - val_loss: 9.8102e-07 Epoch 707/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 5.8934e-06 - val_loss: 4.6523e-06 Epoch 708/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 7.1316e-06 - val_loss: 7.0721e-07 Epoch 709/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 6.8110e-06 - val_loss: 1.3486e-04 Epoch 710/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.8879e-06 - val_loss: 7.0668e-07 Epoch 711/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.0105e-05 - val_loss: 5.9505e-07 Epoch 712/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 8.6121e-07 - val_loss: 1.1204e-06 Epoch 713/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.8833e-06 - val_loss: 1.2931e-06 Epoch 714/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.8541e-06 - val_loss: 1.9471e-06 Epoch 715/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0477e-05 - val_loss: 1.1235e-06 Epoch 716/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0810e-06 - val_loss: 1.5658e-05 Epoch 717/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.0205e-06 - val_loss: 3.8845e-06 Epoch 718/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0286e-05 - val_loss: 3.3220e-05 Epoch 719/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.3179e-06 - val_loss: 7.3284e-07 Epoch 720/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.1861e-06 - val_loss: 5.1867e-05 Epoch 721/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.7525e-05 - val_loss: 2.2818e-06 Epoch 722/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.8037e-06 - val_loss: 1.5087e-06 Epoch 723/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.3043e-06 - val_loss: 1.1848e-06 Epoch 724/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.0499e-06 - val_loss: 1.0393e-06 Epoch 725/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 9.0199e-07 - val_loss: 9.5415e-07 Epoch 726/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.1322e-07 - val_loss: 7.6777e-07 Epoch 727/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.8133e-06 - val_loss: 8.0737e-07 Epoch 728/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 5.4747e-06 - val_loss: 6.0281e-07 Epoch 729/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 8.4922e-06 - val_loss: 7.0930e-07 Epoch 730/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.7676e-06 - val_loss: 1.4913e-05 Epoch 731/1000 3888/3888 [==============================] - 1s 182us/sample - loss: 1.4925e-05 - val_loss: 7.0219e-07 Epoch 732/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 6.4842e-07 - val_loss: 6.2637e-07 Epoch 733/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.7327e-06 - val_loss: 2.0911e-06 Epoch 734/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 8.2876e-07 - val_loss: 1.6520e-06 Epoch 735/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 6.7580e-06 - val_loss: 6.3923e-07 Epoch 736/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.2587e-06 - val_loss: 5.6521e-07 Epoch 737/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 6.0941e-06 - val_loss: 1.8331e-06 Epoch 738/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.6363e-06 - val_loss: 2.8624e-06 Epoch 739/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.9162e-06 - val_loss: 9.1887e-07 Epoch 740/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 4.8136e-06 - val_loss: 2.8889e-06 Epoch 741/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0219e-05 - val_loss: 1.5179e-06 Epoch 742/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.6289e-06 - val_loss: 9.6012e-07 Epoch 743/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.5600e-06 - val_loss: 1.2168e-05 Epoch 744/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.7416e-05 - val_loss: 7.4877e-07 Epoch 745/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.9651e-06 - val_loss: 1.2680e-06 Epoch 746/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.7443e-06 - val_loss: 9.3088e-07 Epoch 747/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.9264e-06 - val_loss: 9.1670e-07 Epoch 748/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.3574e-05 - val_loss: 1.4566e-06 Epoch 749/1000 3888/3888 [==============================] - 1s 181us/sample - loss: 4.5306e-06 - val_loss: 4.7483e-05 Epoch 750/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.8890e-06 - val_loss: 9.5971e-07 Epoch 751/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.9582e-06 - val_loss: 2.1464e-06 Epoch 752/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.9634e-06 - val_loss: 8.4306e-06 Epoch 753/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.6928e-06 - val_loss: 1.6245e-06 Epoch 754/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.1763e-06 - val_loss: 3.2276e-06 Epoch 755/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 9.2350e-06 - val_loss: 2.9095e-05 Epoch 756/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2655e-05 - val_loss: 6.7576e-07 Epoch 757/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.3465e-07 - val_loss: 7.1135e-07 Epoch 758/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.2059e-06 - val_loss: 1.5243e-05 Epoch 759/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.8060e-06 - val_loss: 6.1481e-07 Epoch 760/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.7133e-06 - val_loss: 1.6449e-06 Epoch 761/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 7.6168e-06 - val_loss: 6.3043e-07 Epoch 762/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.8273e-06 - val_loss: 6.3771e-07 Epoch 763/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 6.4535e-06 - val_loss: 3.9212e-06 Epoch 764/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.3539e-06 - val_loss: 1.6992e-06 Epoch 765/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.8220e-05 - val_loss: 1.5792e-05 Epoch 766/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.7985e-06 - val_loss: 6.0404e-07 Epoch 767/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 6.0610e-07 - val_loss: 5.8896e-07 Epoch 768/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.3525e-06 - val_loss: 7.7699e-07 Epoch 769/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.6050e-05 - val_loss: 8.5060e-07 Epoch 770/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 7.1942e-07 - val_loss: 2.7657e-06 Epoch 771/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.3401e-06 - val_loss: 5.6950e-07 Epoch 772/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.5105e-05 - val_loss: 1.8599e-06 Epoch 773/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 8.4921e-07 - val_loss: 1.0381e-06 Epoch 774/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0297e-06 - val_loss: 1.8987e-06 Epoch 775/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 8.3791e-06 - val_loss: 2.9625e-06 Epoch 776/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.8342e-06 - val_loss: 6.7098e-07 Epoch 777/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.9755e-06 - val_loss: 7.7628e-06 Epoch 778/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.1838e-05 - val_loss: 9.9436e-07 Epoch 779/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 9.3956e-07 - val_loss: 1.1311e-06 Epoch 780/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2513e-06 - val_loss: 2.9211e-06 Epoch 781/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.6201e-06 - val_loss: 5.8133e-05 Epoch 782/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 5.7523e-06 - val_loss: 6.5833e-07 Epoch 783/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.3883e-06 - val_loss: 5.7518e-07 Epoch 784/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.3109e-05 - val_loss: 6.4882e-07 Epoch 785/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 9.1963e-07 - val_loss: 7.7708e-07 Epoch 786/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.6003e-06 - val_loss: 9.2990e-07 Epoch 787/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.4349e-05 - val_loss: 1.6265e-06 Epoch 788/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0615e-06 - val_loss: 8.6446e-07 Epoch 789/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.2021e-07 - val_loss: 7.3159e-07 Epoch 790/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.5446e-07 - val_loss: 1.2184e-06 Epoch 791/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.8326e-06 - val_loss: 1.1649e-06 Epoch 792/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 6.5281e-06 - val_loss: 5.2580e-05 Epoch 793/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.8588e-06 - val_loss: 1.9291e-06 Epoch 794/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 9.9394e-06 - val_loss: 1.0689e-06 Epoch 795/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 6.7191e-07 - val_loss: 6.1267e-07 Epoch 796/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.0218e-06 - val_loss: 8.6557e-07 Epoch 797/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.2108e-05 - val_loss: 1.0201e-04 Epoch 798/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.2425e-05 - val_loss: 7.2808e-07 Epoch 799/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 7.0304e-07 - val_loss: 9.6760e-07 Epoch 800/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 7.1558e-07 - val_loss: 5.5140e-06 Epoch 801/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.0549e-06 - val_loss: 1.0227e-05 Epoch 802/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.0821e-05 - val_loss: 1.1289e-06 Epoch 803/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 4.2101e-06 - val_loss: 6.4077e-07 Epoch 804/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 6.6842e-07 - val_loss: 1.3499e-06 Epoch 805/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 5.2447e-06 - val_loss: 9.8721e-07 Epoch 806/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 7.1325e-06 - val_loss: 1.0566e-04 Epoch 807/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.3198e-05 - val_loss: 9.1788e-07 Epoch 808/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0705e-06 - val_loss: 9.6235e-07 Epoch 809/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 6.5115e-06 - val_loss: 1.7538e-05 Epoch 810/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.8441e-05 - val_loss: 9.0477e-07 Epoch 811/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 8.3060e-07 - val_loss: 7.6046e-07 Epoch 812/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 8.3633e-07 - val_loss: 8.1277e-07 Epoch 813/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.1468e-05 - val_loss: 7.3523e-07 Epoch 814/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 7.0558e-07 - val_loss: 6.6015e-07 Epoch 815/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.1218e-06 - val_loss: 1.4832e-05 Epoch 816/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.3783e-06 - val_loss: 7.2879e-07 Epoch 817/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.5321e-06 - val_loss: 5.8072e-06 Epoch 818/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 4.2149e-06 - val_loss: 2.2895e-05 Epoch 819/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.5535e-06 - val_loss: 4.5771e-06 Epoch 820/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.0600e-05 - val_loss: 1.6808e-06 Epoch 821/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.2665e-06 - val_loss: 1.2115e-06 Epoch 822/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.0530e-06 - val_loss: 9.3358e-07 Epoch 823/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.9179e-07 - val_loss: 8.2348e-07 Epoch 824/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 7.9828e-07 - val_loss: 7.9668e-07 Epoch 825/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.7851e-06 - val_loss: 7.2348e-07 Epoch 826/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.5011e-06 - val_loss: 7.4602e-07 Epoch 827/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.4603e-06 - val_loss: 1.5703e-06 Epoch 828/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0309e-05 - val_loss: 1.0815e-05 Epoch 829/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2254e-06 - val_loss: 3.3511e-06 Epoch 830/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.3901e-06 - val_loss: 5.2407e-05 Epoch 831/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.9628e-06 - val_loss: 2.0404e-06 Epoch 832/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.4720e-06 - val_loss: 9.3301e-07 Epoch 833/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.7215e-06 - val_loss: 3.0003e-06 Epoch 834/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.0808e-05 - val_loss: 7.2190e-07 Epoch 835/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.8221e-07 - val_loss: 7.6455e-07 Epoch 836/1000 3840/3888 [============================>.] - ETA: 0s - loss: 1.5360e-06Restoring model weights from the end of the best epoch. 3888/3888 [==============================] - 1s 177us/sample - loss: 1.6198e-06 - val_loss: 1.0451e-05 Epoch 00836: early stopping
In [39]:
print(history.history.keys())
print('best value: ', autoencoder.evaluate(X_train_1D_norm, X_train_1D_norm, verbose=0))
pd.DataFrame(history.history).plot(figsize=(8, 5), logy=True)
plt.grid()
dict_keys(['loss', 'val_loss']) best value: 5.652084395026656e-07
In [40]:
X_reconstructions = autoencoder.predict(X_train_1D_norm)
X_reconstructions = stdscaler.inverse_transform(X_reconstructions)
In [41]:
calculateerror(X_train_1D.reshape(len(times),len(groups),nl,nc),
X_reconstructions.reshape(len(times),len(groups),nl,nc),
groups,
print_step=0)
max_abs_error: 5.275390625 mean_abs_error: 0.015190057506593278
/home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: divide by zero encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until /home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until
In [42]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(X_reconstructions.reshape(len(times),len(groups),nl,nc)[100,i,:,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [43]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_train_1D[:,i*nl*nc+4])
ax.flatten()[i].plot(times, X_reconstructions[:,i*nl*nc+4],'--')
ax.flatten()[i].set_title(group)
Non-linear autoencoder¶
In [44]:
np.random.seed(42)
tf.random.set_seed(42)
# Need to have validation loss
early_stopping = keras.callbacks.EarlyStopping(monitor='val_loss',
min_delta=0.0,
patience=100,
verbose=2,
restore_best_weights=True)
encoder = keras.models.Sequential([keras.layers.Dense(100, input_shape=[800], activation="elu"),
keras.layers.Dense(50, activation="elu"),
keras.layers.Dense(15)])
decoder = keras.models.Sequential([keras.layers.Dense(50, input_shape=[15], activation="elu"),
keras.layers.Dense(100, activation="elu"),
keras.layers.Dense(800),
])
autoencoder = keras.models.Sequential([encoder, decoder])
autoencoder.compile(loss="mse",
optimizer=keras.optimizers.Nadam(lr=0.0003, beta_1=0.9, beta_2=0.999)
)
encoder.summary()
decoder.summary()
Model: "sequential_6" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= dense_6 (Dense) (None, 100) 80100 _________________________________________________________________ dense_7 (Dense) (None, 50) 5050 _________________________________________________________________ dense_8 (Dense) (None, 15) 765 ================================================================= Total params: 85,915 Trainable params: 85,915 Non-trainable params: 0 _________________________________________________________________ Model: "sequential_7" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= dense_9 (Dense) (None, 50) 800 _________________________________________________________________ dense_10 (Dense) (None, 100) 5100 _________________________________________________________________ dense_11 (Dense) (None, 800) 80800 ================================================================= Total params: 86,700 Trainable params: 86,700 Non-trainable params: 0 _________________________________________________________________
In [45]:
history = autoencoder.fit(X_train_1D_norm,
X_train_1D_norm,
epochs=1000,
validation_data=(X_train_1D_norm, X_train_1D_norm),
callbacks=[early_stopping])
Train on 3888 samples, validate on 3888 samples Epoch 1/1000 3888/3888 [==============================] - 2s 399us/sample - loss: 0.0738 - val_loss: 0.0213 Epoch 2/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 0.0133 - val_loss: 0.0077 Epoch 3/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 0.0052 - val_loss: 0.0039 Epoch 4/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 0.0028 - val_loss: 0.0019 Epoch 5/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 0.0016 - val_loss: 0.0014 Epoch 6/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 0.0011 - val_loss: 9.0248e-04 Epoch 7/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 8.9075e-04 - val_loss: 8.0369e-04 Epoch 8/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 6.9541e-04 - val_loss: 5.6254e-04 Epoch 9/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 6.2126e-04 - val_loss: 4.4743e-04 Epoch 10/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 4.3809e-04 - val_loss: 4.5001e-04 Epoch 11/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.3138e-04 - val_loss: 3.2076e-04 Epoch 12/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.2502e-04 - val_loss: 3.9755e-04 Epoch 13/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.5841e-04 - val_loss: 2.4384e-04 Epoch 14/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.9576e-04 - val_loss: 2.2604e-04 Epoch 15/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.6657e-04 - val_loss: 2.0372e-04 Epoch 16/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.3324e-04 - val_loss: 1.9059e-04 Epoch 17/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.2792e-04 - val_loss: 1.8347e-04 Epoch 18/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.7011e-04 - val_loss: 2.6279e-04 Epoch 19/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.4473e-04 - val_loss: 1.5079e-04 Epoch 20/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.1813e-04 - val_loss: 1.3292e-04 Epoch 21/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.4057e-04 - val_loss: 1.1185e-04 Epoch 22/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.6320e-04 - val_loss: 1.2385e-04 Epoch 23/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.6112e-04 - val_loss: 1.1233e-04 Epoch 24/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.2680e-04 - val_loss: 1.2117e-04 Epoch 25/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.0080e-04 - val_loss: 8.9527e-05 Epoch 26/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.8405e-04 - val_loss: 2.4367e-04 Epoch 27/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0457e-04 - val_loss: 1.6173e-04 Epoch 28/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.2771e-04 - val_loss: 1.1469e-04 Epoch 29/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.1713e-04 - val_loss: 3.4487e-04 Epoch 30/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.1976e-04 - val_loss: 3.4356e-04 Epoch 31/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.0136e-04 - val_loss: 6.7224e-05 Epoch 32/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 7.7168e-05 - val_loss: 8.2778e-05 Epoch 33/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.4334e-04 - val_loss: 6.7871e-05 Epoch 34/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.6404e-05 - val_loss: 6.4198e-05 Epoch 35/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 7.6001e-05 - val_loss: 1.2657e-04 Epoch 36/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.1915e-04 - val_loss: 5.4725e-05 Epoch 37/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.2745e-05 - val_loss: 7.7216e-05 Epoch 38/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.1707e-04 - val_loss: 6.1423e-05 Epoch 39/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 7.0213e-05 - val_loss: 5.9363e-05 Epoch 40/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.8709e-04 - val_loss: 5.8366e-05 Epoch 41/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.0724e-05 - val_loss: 4.5737e-05 Epoch 42/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.0222e-05 - val_loss: 8.8702e-05 Epoch 43/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 6.3829e-05 - val_loss: 8.8226e-05 Epoch 44/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.6035e-05 - val_loss: 6.1917e-05 Epoch 45/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 5.9398e-05 - val_loss: 1.4935e-04 Epoch 46/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0370e-04 - val_loss: 3.9682e-05 Epoch 47/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 7.8800e-05 - val_loss: 4.5692e-05 Epoch 48/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 4.5110e-05 - val_loss: 3.9648e-05 Epoch 49/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.2317e-04 - val_loss: 3.4771e-05 Epoch 50/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.2581e-05 - val_loss: 5.9154e-05 Epoch 51/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.7513e-05 - val_loss: 3.4424e-05 Epoch 52/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.1373e-05 - val_loss: 3.7781e-05 Epoch 53/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 6.2329e-05 - val_loss: 3.3030e-05 Epoch 54/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 9.8988e-05 - val_loss: 3.8013e-05 Epoch 55/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 3.4162e-05 - val_loss: 3.0226e-05 Epoch 56/1000 3888/3888 [==============================] - 1s 160us/sample - loss: 3.5709e-05 - val_loss: 4.9157e-05 Epoch 57/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 5.7291e-05 - val_loss: 4.9004e-05 Epoch 58/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 6.7873e-05 - val_loss: 6.6344e-04 Epoch 59/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.6349e-05 - val_loss: 4.0357e-05 Epoch 60/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.7941e-05 - val_loss: 2.5937e-05 Epoch 61/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.5275e-05 - val_loss: 3.8480e-05 Epoch 62/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 9.3105e-05 - val_loss: 2.7316e-05 Epoch 63/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.0758e-05 - val_loss: 2.5589e-05 Epoch 64/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.1119e-05 - val_loss: 2.4874e-05 Epoch 65/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 3.4365e-05 - val_loss: 6.7826e-05 Epoch 66/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 5.0768e-05 - val_loss: 2.5311e-05 Epoch 67/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 4.2197e-05 - val_loss: 2.6397e-04 Epoch 68/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.9717e-05 - val_loss: 2.2434e-05 Epoch 69/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.6274e-05 - val_loss: 3.1011e-05 Epoch 70/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.2399e-05 - val_loss: 2.5650e-05 Epoch 71/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.5326e-05 - val_loss: 2.0978e-05 Epoch 72/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.6382e-05 - val_loss: 2.2463e-05 Epoch 73/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 5.5984e-05 - val_loss: 2.4758e-05 Epoch 74/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.9602e-05 - val_loss: 2.2974e-05 Epoch 75/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.0904e-05 - val_loss: 2.1345e-05 Epoch 76/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.3756e-05 - val_loss: 2.2046e-05 Epoch 77/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 4.9396e-05 - val_loss: 5.9935e-05 Epoch 78/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.3136e-05 - val_loss: 2.0602e-05 Epoch 79/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.9133e-05 - val_loss: 2.5864e-04 Epoch 80/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.3918e-05 - val_loss: 2.1660e-05 Epoch 81/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.1656e-05 - val_loss: 4.1986e-05 Epoch 82/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.0388e-05 - val_loss: 2.4040e-05 Epoch 83/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.8229e-05 - val_loss: 1.5414e-05 Epoch 84/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.3026e-05 - val_loss: 2.1708e-05 Epoch 85/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.8399e-05 - val_loss: 9.3783e-05 Epoch 86/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.3101e-05 - val_loss: 3.2621e-05 Epoch 87/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.1309e-05 - val_loss: 3.1538e-05 Epoch 88/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.8559e-05 - val_loss: 6.7896e-05 Epoch 89/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.1038e-05 - val_loss: 2.5238e-05 Epoch 90/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.9451e-05 - val_loss: 5.2010e-05 Epoch 91/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 3.9412e-05 - val_loss: 1.6047e-05 Epoch 92/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.8037e-05 - val_loss: 3.8796e-05 Epoch 93/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.2766e-05 - val_loss: 1.5509e-05 Epoch 94/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.2778e-05 - val_loss: 2.4701e-05 Epoch 95/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.7686e-05 - val_loss: 3.5350e-05 Epoch 96/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.3565e-05 - val_loss: 3.4616e-05 Epoch 97/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.0037e-05 - val_loss: 5.3302e-05 Epoch 98/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.4685e-05 - val_loss: 1.4850e-05 Epoch 99/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.0763e-05 - val_loss: 1.8319e-05 Epoch 100/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.9349e-05 - val_loss: 1.6575e-05 Epoch 101/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.8836e-05 - val_loss: 2.0057e-05 Epoch 102/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.5714e-05 - val_loss: 1.5270e-05 Epoch 103/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.9804e-05 - val_loss: 3.0410e-05 Epoch 104/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.4888e-05 - val_loss: 2.2791e-05 Epoch 105/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.9463e-05 - val_loss: 2.6771e-05 Epoch 106/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.7890e-05 - val_loss: 1.2698e-05 Epoch 107/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.5256e-05 - val_loss: 1.1501e-05 Epoch 108/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.7900e-05 - val_loss: 1.1755e-05 Epoch 109/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.9958e-05 - val_loss: 1.3093e-05 Epoch 110/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.4672e-05 - val_loss: 1.1263e-05 Epoch 111/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.2315e-05 - val_loss: 1.3526e-05 Epoch 112/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.0585e-05 - val_loss: 2.0821e-05 Epoch 113/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.4648e-05 - val_loss: 1.1934e-05 Epoch 114/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.9286e-05 - val_loss: 3.2732e-05 Epoch 115/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.6027e-05 - val_loss: 2.0702e-05 Epoch 116/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.1921e-05 - val_loss: 7.7981e-05 Epoch 117/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.5643e-05 - val_loss: 1.1578e-05 Epoch 118/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.1226e-05 - val_loss: 1.7894e-05 Epoch 119/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.0202e-05 - val_loss: 1.5657e-05 Epoch 120/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.2040e-05 - val_loss: 1.2324e-05 Epoch 121/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.0070e-05 - val_loss: 2.6288e-04 Epoch 122/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.2631e-05 - val_loss: 1.1904e-05 Epoch 123/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.7329e-05 - val_loss: 1.5348e-05 Epoch 124/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.3928e-05 - val_loss: 1.2670e-05 Epoch 125/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.0412e-05 - val_loss: 9.3581e-06 Epoch 126/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.1771e-05 - val_loss: 1.8831e-05 Epoch 127/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.5334e-05 - val_loss: 4.7775e-05 Epoch 128/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.2387e-05 - val_loss: 1.7573e-05 Epoch 129/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.6525e-05 - val_loss: 1.2063e-05 Epoch 130/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0571e-05 - val_loss: 1.3839e-05 Epoch 131/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.3311e-05 - val_loss: 1.2631e-05 Epoch 132/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.2047e-05 - val_loss: 9.7373e-06 Epoch 133/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.7530e-05 - val_loss: 3.6506e-05 Epoch 134/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.5740e-05 - val_loss: 9.0237e-06 Epoch 135/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.9252e-06 - val_loss: 2.2183e-05 Epoch 136/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.6348e-05 - val_loss: 1.7399e-05 Epoch 137/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0603e-05 - val_loss: 2.2265e-05 Epoch 138/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.2988e-05 - val_loss: 1.6963e-05 Epoch 139/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.4670e-05 - val_loss: 1.0630e-05 Epoch 140/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.1128e-05 - val_loss: 1.2292e-05 Epoch 141/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.5385e-05 - val_loss: 2.9868e-05 Epoch 142/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.9851e-05 - val_loss: 7.6551e-06 Epoch 143/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.8339e-05 - val_loss: 7.6110e-06 Epoch 144/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.5839e-06 - val_loss: 9.5788e-06 Epoch 145/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.5811e-05 - val_loss: 8.3010e-06 Epoch 146/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0925e-05 - val_loss: 9.0989e-06 Epoch 147/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.8010e-05 - val_loss: 1.3611e-05 Epoch 148/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.8578e-05 - val_loss: 6.2824e-05 Epoch 149/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.5082e-05 - val_loss: 7.4335e-06 Epoch 150/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.2927e-05 - val_loss: 1.1447e-05 Epoch 151/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 8.4998e-06 - val_loss: 9.4952e-06 Epoch 152/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.0168e-05 - val_loss: 1.2642e-05 Epoch 153/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.8704e-05 - val_loss: 8.0965e-06 Epoch 154/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.5729e-05 - val_loss: 6.6933e-05 Epoch 155/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.3284e-05 - val_loss: 8.4163e-06 Epoch 156/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 7.5732e-06 - val_loss: 7.9060e-06 Epoch 157/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.4612e-05 - val_loss: 7.6637e-06 Epoch 158/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0905e-05 - val_loss: 7.8848e-06 Epoch 159/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.0838e-05 - val_loss: 1.2462e-05 Epoch 160/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0337e-05 - val_loss: 6.2857e-05 Epoch 161/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.9129e-05 - val_loss: 1.1350e-05 Epoch 162/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.3920e-05 - val_loss: 1.4463e-05 Epoch 163/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.7938e-05 - val_loss: 6.8174e-06 Epoch 164/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.9587e-05 - val_loss: 7.8949e-06 Epoch 165/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 8.5997e-06 - val_loss: 1.5012e-05 Epoch 166/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 6.2674e-05 - val_loss: 6.8490e-06 Epoch 167/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.6387e-06 - val_loss: 1.4120e-05 Epoch 168/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 7.4868e-06 - val_loss: 8.3199e-06 Epoch 169/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.8242e-06 - val_loss: 8.4830e-06 Epoch 170/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.1194e-05 - val_loss: 1.5911e-05 Epoch 171/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2357e-05 - val_loss: 1.3681e-05 Epoch 172/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.8332e-05 - val_loss: 9.6728e-06 Epoch 173/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.2213e-05 - val_loss: 9.6931e-06 Epoch 174/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.3268e-05 - val_loss: 1.3382e-05 Epoch 175/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.5114e-05 - val_loss: 2.7774e-05 Epoch 176/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 7.2065e-06 - val_loss: 1.6149e-05 Epoch 177/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.4761e-05 - val_loss: 1.2932e-05 Epoch 178/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.8781e-06 - val_loss: 1.8651e-05 Epoch 179/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0437e-05 - val_loss: 1.8698e-05 Epoch 180/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.2197e-05 - val_loss: 6.6164e-06 Epoch 181/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 9.9575e-06 - val_loss: 8.0228e-06 Epoch 182/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.6886e-05 - val_loss: 7.5428e-06 Epoch 183/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.0089e-05 - val_loss: 2.7721e-05 Epoch 184/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.2796e-05 - val_loss: 6.6397e-06 Epoch 185/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.6497e-05 - val_loss: 1.2537e-05 Epoch 186/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.1523e-05 - val_loss: 7.4342e-06 Epoch 187/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 2.3104e-05 - val_loss: 7.4411e-06 Epoch 188/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.1632e-05 - val_loss: 7.5084e-06 Epoch 189/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.4408e-05 - val_loss: 9.8039e-06 Epoch 190/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.0078e-05 - val_loss: 4.7622e-05 Epoch 191/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.3345e-05 - val_loss: 1.0189e-05 Epoch 192/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.5226e-05 - val_loss: 1.3804e-05 Epoch 193/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.8407e-05 - val_loss: 6.1693e-06 Epoch 194/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0482e-05 - val_loss: 6.0754e-06 Epoch 195/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.7052e-05 - val_loss: 1.0908e-05 Epoch 196/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.2090e-05 - val_loss: 6.7758e-06 Epoch 197/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.9872e-05 - val_loss: 5.4886e-06 Epoch 198/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 9.1041e-06 - val_loss: 1.6037e-05 Epoch 199/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.3477e-05 - val_loss: 7.5185e-06 Epoch 200/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.6129e-06 - val_loss: 6.1187e-06 Epoch 201/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.3053e-06 - val_loss: 9.7243e-06 Epoch 202/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.1907e-06 - val_loss: 5.7320e-06 Epoch 203/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.0559e-05 - val_loss: 7.0414e-06 Epoch 204/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.3038e-06 - val_loss: 5.5673e-06 Epoch 205/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.3877e-05 - val_loss: 5.4835e-06 Epoch 206/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 1.4762e-05 - val_loss: 8.2328e-06 Epoch 207/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 8.8227e-06 - val_loss: 1.1109e-05 Epoch 208/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.0193e-05 - val_loss: 1.0757e-05 Epoch 209/1000 3888/3888 [==============================] - 1s 159us/sample - loss: 1.6210e-05 - val_loss: 2.2095e-05 Epoch 210/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.0547e-05 - val_loss: 1.2884e-05 Epoch 211/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.0618e-05 - val_loss: 1.3064e-04 Epoch 212/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.9069e-05 - val_loss: 5.1850e-06 Epoch 213/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 9.1538e-06 - val_loss: 5.5521e-06 Epoch 214/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.0057e-05 - val_loss: 7.6671e-06 Epoch 215/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.5276e-06 - val_loss: 9.1512e-06 Epoch 216/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.5359e-05 - val_loss: 6.9617e-06 Epoch 217/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.2081e-05 - val_loss: 5.0828e-06 Epoch 218/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 7.0452e-06 - val_loss: 5.1056e-06 Epoch 219/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.5939e-05 - val_loss: 3.0916e-05 Epoch 220/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 7.5386e-06 - val_loss: 1.8468e-05 Epoch 221/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 6.7852e-06 - val_loss: 5.2979e-06 Epoch 222/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.1551e-05 - val_loss: 7.2192e-06 Epoch 223/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0054e-05 - val_loss: 9.7108e-06 Epoch 224/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.7802e-06 - val_loss: 5.2869e-06 Epoch 225/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.3071e-05 - val_loss: 1.2655e-05 Epoch 226/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 8.8989e-06 - val_loss: 1.2160e-05 Epoch 227/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.4125e-05 - val_loss: 5.1797e-06 Epoch 228/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0808e-05 - val_loss: 2.1585e-05 Epoch 229/1000 3888/3888 [==============================] - 1s 159us/sample - loss: 1.3656e-05 - val_loss: 1.2228e-05 Epoch 230/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.3534e-05 - val_loss: 6.0582e-06 Epoch 231/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.3763e-06 - val_loss: 6.1127e-06 Epoch 232/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0561e-05 - val_loss: 6.4672e-06 Epoch 233/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.1756e-06 - val_loss: 2.3157e-05 Epoch 234/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0226e-05 - val_loss: 5.3701e-06 Epoch 235/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.1181e-05 - val_loss: 6.0515e-06 Epoch 236/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.5069e-05 - val_loss: 2.3999e-04 Epoch 237/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.8905e-05 - val_loss: 4.2783e-06 Epoch 238/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.0404e-06 - val_loss: 4.8141e-06 Epoch 239/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 8.7867e-06 - val_loss: 1.2584e-04 Epoch 240/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.0431e-05 - val_loss: 4.4344e-06 Epoch 241/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.7971e-06 - val_loss: 1.1865e-05 Epoch 242/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.2251e-06 - val_loss: 4.8635e-05 Epoch 243/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 7.8126e-06 - val_loss: 1.9679e-05 Epoch 244/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.5146e-05 - val_loss: 5.3746e-06 Epoch 245/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.7724e-05 - val_loss: 6.1827e-06 Epoch 246/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 7.7808e-06 - val_loss: 6.7323e-06 Epoch 247/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.7133e-06 - val_loss: 5.1471e-06 Epoch 248/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.5327e-05 - val_loss: 1.7223e-05 Epoch 249/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 9.5306e-06 - val_loss: 9.6417e-05 Epoch 250/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.4026e-05 - val_loss: 5.1316e-06 Epoch 251/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.3344e-06 - val_loss: 3.5736e-06 Epoch 252/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.9748e-06 - val_loss: 4.2490e-06 Epoch 253/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.9845e-06 - val_loss: 4.9721e-06 Epoch 254/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.4673e-06 - val_loss: 8.0273e-05 Epoch 255/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.8333e-06 - val_loss: 4.5998e-06 Epoch 256/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 9.8041e-06 - val_loss: 3.0341e-05 Epoch 257/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.1863e-05 - val_loss: 2.3186e-05 Epoch 258/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.1555e-05 - val_loss: 5.4273e-06 Epoch 259/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.1461e-06 - val_loss: 5.3083e-06 Epoch 260/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.2189e-05 - val_loss: 4.7553e-06 Epoch 261/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 8.7927e-06 - val_loss: 6.1851e-06 Epoch 262/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0477e-05 - val_loss: 5.3305e-06 Epoch 263/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0374e-05 - val_loss: 6.2210e-06 Epoch 264/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.6051e-05 - val_loss: 1.0271e-05 Epoch 265/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.1601e-05 - val_loss: 4.6802e-06 Epoch 266/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.3733e-05 - val_loss: 4.6581e-06 Epoch 267/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.4845e-06 - val_loss: 7.0475e-06 Epoch 268/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0336e-05 - val_loss: 1.4583e-05 Epoch 269/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 7.9788e-06 - val_loss: 1.2923e-04 Epoch 270/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.4877e-05 - val_loss: 4.0521e-06 Epoch 271/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.7689e-05 - val_loss: 5.7481e-06 Epoch 272/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.6589e-06 - val_loss: 4.3783e-06 Epoch 273/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 5.9236e-06 - val_loss: 4.5348e-05 Epoch 274/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.0428e-05 - val_loss: 8.1480e-06 Epoch 275/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 9.4357e-06 - val_loss: 9.4461e-06 Epoch 276/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.7846e-05 - val_loss: 1.1419e-05 Epoch 277/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 6.7671e-06 - val_loss: 6.2370e-06 Epoch 278/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.4216e-06 - val_loss: 4.7033e-06 Epoch 279/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.4102e-06 - val_loss: 6.3642e-06 Epoch 280/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.2950e-05 - val_loss: 9.8561e-06 Epoch 281/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 6.4483e-06 - val_loss: 5.0064e-06 Epoch 282/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.0415e-05 - val_loss: 3.8922e-06 Epoch 283/1000 3888/3888 [==============================] - 1s 157us/sample - loss: 7.3388e-06 - val_loss: 2.8215e-05 Epoch 284/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.9822e-05 - val_loss: 4.6206e-06 Epoch 285/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 8.4450e-06 - val_loss: 6.4214e-06 Epoch 286/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.0932e-06 - val_loss: 9.2331e-06 Epoch 287/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.1163e-05 - val_loss: 3.6472e-06 Epoch 288/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.8076e-05 - val_loss: 1.8491e-05 Epoch 289/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 6.2439e-06 - val_loss: 2.0756e-05 Epoch 290/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 8.1149e-06 - val_loss: 5.9633e-06 Epoch 291/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.0015e-05 - val_loss: 8.8248e-06 Epoch 292/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0581e-05 - val_loss: 7.4719e-06 Epoch 293/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 9.3946e-06 - val_loss: 1.1636e-04 Epoch 294/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.3211e-05 - val_loss: 5.5047e-06 Epoch 295/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.3807e-05 - val_loss: 1.3238e-05 Epoch 296/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.7157e-05 - val_loss: 1.2660e-05 Epoch 297/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.3122e-06 - val_loss: 5.9720e-06 Epoch 298/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.9270e-06 - val_loss: 5.7431e-06 Epoch 299/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.6391e-05 - val_loss: 1.0898e-05 Epoch 300/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 5.5745e-06 - val_loss: 3.5126e-06 Epoch 301/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.4149e-05 - val_loss: 2.5050e-05 Epoch 302/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.1808e-05 - val_loss: 4.6985e-06 Epoch 303/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 8.0499e-06 - val_loss: 3.0485e-05 Epoch 304/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.1508e-05 - val_loss: 5.7677e-06 Epoch 305/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.4347e-05 - val_loss: 4.1689e-06 Epoch 306/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.5567e-06 - val_loss: 1.1056e-05 Epoch 307/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 7.4921e-06 - val_loss: 3.2717e-06 Epoch 308/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.1503e-05 - val_loss: 4.2534e-06 Epoch 309/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.5902e-06 - val_loss: 7.8330e-06 Epoch 310/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.3964e-06 - val_loss: 4.3555e-06 Epoch 311/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.2975e-05 - val_loss: 1.5889e-05 Epoch 312/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.2725e-06 - val_loss: 3.4147e-06 Epoch 313/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.9557e-06 - val_loss: 5.5331e-06 Epoch 314/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.0954e-06 - val_loss: 4.6023e-06 Epoch 315/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.2711e-06 - val_loss: 1.2843e-05 Epoch 316/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 6.0468e-06 - val_loss: 5.5753e-06 Epoch 317/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.2357e-05 - val_loss: 1.0158e-05 Epoch 318/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.8644e-06 - val_loss: 3.6753e-06 Epoch 319/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.1166e-05 - val_loss: 1.6105e-05 Epoch 320/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 6.7154e-06 - val_loss: 4.0704e-06 Epoch 321/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.1132e-05 - val_loss: 5.2871e-06 Epoch 322/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 9.2533e-06 - val_loss: 7.7238e-06 Epoch 323/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.4442e-05 - val_loss: 5.3351e-06 Epoch 324/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.4400e-06 - val_loss: 4.7407e-06 Epoch 325/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 5.3865e-06 - val_loss: 4.7225e-06 Epoch 326/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 8.6586e-06 - val_loss: 5.8643e-06 Epoch 327/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.8807e-05 - val_loss: 6.0742e-06 Epoch 328/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.9499e-06 - val_loss: 1.9017e-05 Epoch 329/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.5070e-06 - val_loss: 1.0281e-05 Epoch 330/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.3545e-05 - val_loss: 4.3046e-06 Epoch 331/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.6358e-06 - val_loss: 7.7832e-06 Epoch 332/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.2000e-05 - val_loss: 3.9405e-06 Epoch 333/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 5.3208e-06 - val_loss: 3.6070e-06 Epoch 334/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.1551e-05 - val_loss: 5.1382e-06 Epoch 335/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.8319e-05 - val_loss: 4.1359e-04 Epoch 336/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.1675e-05 - val_loss: 3.3741e-06 Epoch 337/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.7597e-06 - val_loss: 4.8394e-06 Epoch 338/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.9801e-06 - val_loss: 9.6883e-06 Epoch 339/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.5459e-06 - val_loss: 5.6855e-06 Epoch 340/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.1239e-05 - val_loss: 4.5398e-06 Epoch 341/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.5917e-06 - val_loss: 3.9320e-06 Epoch 342/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 9.7760e-06 - val_loss: 2.3736e-04 Epoch 343/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 8.0052e-06 - val_loss: 2.9829e-06 Epoch 344/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0756e-05 - val_loss: 2.9324e-06 Epoch 345/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.2911e-05 - val_loss: 1.4155e-05 Epoch 346/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.0821e-06 - val_loss: 3.7546e-06 Epoch 347/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 6.4453e-06 - val_loss: 3.6433e-06 Epoch 348/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 6.6838e-06 - val_loss: 9.2481e-05 Epoch 349/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 7.7622e-06 - val_loss: 3.6064e-06 Epoch 350/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 8.4285e-06 - val_loss: 2.6588e-05 Epoch 351/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.1593e-05 - val_loss: 3.9792e-06 Epoch 352/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.3230e-05 - val_loss: 5.6334e-06 Epoch 353/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.7762e-06 - val_loss: 3.5595e-06 Epoch 354/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 6.2687e-06 - val_loss: 7.4060e-06 Epoch 355/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0162e-05 - val_loss: 1.0944e-05 Epoch 356/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.7119e-05 - val_loss: 5.0458e-06 Epoch 357/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.9935e-06 - val_loss: 3.6135e-06 Epoch 358/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.4689e-06 - val_loss: 4.4243e-06 Epoch 359/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.9716e-06 - val_loss: 3.2339e-05 Epoch 360/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.0171e-05 - val_loss: 5.6580e-06 Epoch 361/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 8.6792e-06 - val_loss: 1.2932e-05 Epoch 362/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 5.1457e-06 - val_loss: 5.4555e-06 Epoch 363/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 8.6839e-06 - val_loss: 4.0269e-06 Epoch 364/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.8739e-05 - val_loss: 4.0364e-06 Epoch 365/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.9905e-06 - val_loss: 2.7873e-06 Epoch 366/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 2.9259e-06 - val_loss: 2.6409e-06 Epoch 367/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 3.5596e-06 - val_loss: 1.6126e-05 Epoch 368/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.2055e-05 - val_loss: 6.3265e-06 Epoch 369/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.1508e-06 - val_loss: 4.7824e-06 Epoch 370/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 1.2494e-05 - val_loss: 4.8266e-06 Epoch 371/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.9883e-06 - val_loss: 3.4817e-06 Epoch 372/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 3.5770e-06 - val_loss: 3.2407e-06 Epoch 373/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 9.9641e-06 - val_loss: 3.9142e-06 Epoch 374/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 4.4466e-06 - val_loss: 1.0860e-05 Epoch 375/1000 3888/3888 [==============================] - 1s 178us/sample - loss: 1.5610e-05 - val_loss: 5.2315e-06 Epoch 376/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 4.1728e-06 - val_loss: 3.7117e-06 Epoch 377/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 7.6357e-06 - val_loss: 3.1505e-06 Epoch 378/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 7.7787e-06 - val_loss: 6.8963e-06 Epoch 379/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 1.1418e-05 - val_loss: 3.6732e-05 Epoch 380/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.2511e-05 - val_loss: 2.3034e-05 Epoch 381/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 5.6154e-06 - val_loss: 9.4374e-06 Epoch 382/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 8.7741e-06 - val_loss: 3.5170e-06 Epoch 383/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.9549e-05 - val_loss: 1.2875e-05 Epoch 384/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.6103e-06 - val_loss: 2.6301e-06 Epoch 385/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.4067e-06 - val_loss: 5.3911e-06 Epoch 386/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.1253e-06 - val_loss: 4.5703e-06 Epoch 387/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 8.5378e-06 - val_loss: 6.4900e-05 Epoch 388/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.4220e-05 - val_loss: 3.5075e-06 Epoch 389/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.6732e-06 - val_loss: 3.4816e-06 Epoch 390/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 7.8688e-06 - val_loss: 3.3996e-06 Epoch 391/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 7.2387e-06 - val_loss: 2.7627e-06 Epoch 392/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 9.7411e-06 - val_loss: 5.9485e-06 Epoch 393/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 9.2277e-06 - val_loss: 3.3208e-06 Epoch 394/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.3487e-05 - val_loss: 3.4789e-06 Epoch 395/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.8777e-06 - val_loss: 8.1317e-06 Epoch 396/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.5382e-05 - val_loss: 2.7258e-05 Epoch 397/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.9436e-06 - val_loss: 3.2157e-06 Epoch 398/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.3265e-06 - val_loss: 3.2238e-06 Epoch 399/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.2774e-06 - val_loss: 3.1650e-06 Epoch 400/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 7.5438e-06 - val_loss: 1.2491e-05 Epoch 401/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.4453e-06 - val_loss: 3.9355e-06 Epoch 402/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.2062e-06 - val_loss: 5.7513e-06 Epoch 403/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 9.6880e-06 - val_loss: 1.5079e-05 Epoch 404/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.2819e-05 - val_loss: 2.9499e-06 Epoch 405/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.9639e-06 - val_loss: 3.8571e-06 Epoch 406/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.4410e-05 - val_loss: 3.9061e-06 Epoch 407/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.1977e-06 - val_loss: 4.5615e-06 Epoch 408/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 6.6595e-06 - val_loss: 2.6380e-06 Epoch 409/1000 3888/3888 [==============================] - 1s 179us/sample - loss: 3.9393e-06 - val_loss: 4.3143e-05 Epoch 410/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 8.7277e-06 - val_loss: 3.6748e-06 Epoch 411/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 6.0651e-06 - val_loss: 1.1370e-05 Epoch 412/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 5.6923e-06 - val_loss: 8.1826e-06 Epoch 413/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 8.7038e-06 - val_loss: 3.1376e-06 Epoch 414/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.9015e-05 - val_loss: 5.0624e-06 Epoch 415/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 3.9656e-06 - val_loss: 2.2221e-05 Epoch 416/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 9.3168e-06 - val_loss: 7.6860e-06 Epoch 417/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.9901e-05 - val_loss: 3.5244e-06 Epoch 418/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.1827e-06 - val_loss: 3.2904e-06 Epoch 419/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.8657e-06 - val_loss: 1.0701e-05 Epoch 420/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.7142e-06 - val_loss: 3.1441e-06 Epoch 421/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0619e-05 - val_loss: 3.3862e-06 Epoch 422/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.3337e-06 - val_loss: 3.6045e-06 Epoch 423/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.1331e-05 - val_loss: 1.8343e-05 Epoch 424/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.2488e-05 - val_loss: 4.6157e-06 Epoch 425/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.0622e-06 - val_loss: 8.2734e-06 Epoch 426/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.6168e-06 - val_loss: 4.3664e-06 Epoch 427/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.3592e-06 - val_loss: 1.1169e-05 Epoch 428/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.6308e-05 - val_loss: 4.4296e-05 Epoch 429/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.4065e-06 - val_loss: 3.2271e-05 Epoch 430/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.4491e-06 - val_loss: 2.5584e-06 Epoch 431/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 8.9666e-06 - val_loss: 7.0473e-06 Epoch 432/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 8.6746e-06 - val_loss: 5.7882e-06 Epoch 433/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 9.4038e-06 - val_loss: 4.3911e-06 Epoch 434/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 4.8378e-06 - val_loss: 1.0350e-05 Epoch 435/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 8.5769e-06 - val_loss: 7.5321e-06 Epoch 436/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.2446e-05 - val_loss: 7.5101e-06 Epoch 437/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.1076e-05 - val_loss: 4.2544e-06 Epoch 438/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.4073e-06 - val_loss: 9.7141e-06 Epoch 439/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.1809e-05 - val_loss: 4.1786e-06 Epoch 440/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 3.5189e-06 - val_loss: 4.6077e-06 Epoch 441/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 4.0842e-06 - val_loss: 3.7683e-06 Epoch 442/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 5.0600e-06 - val_loss: 1.6969e-05 Epoch 443/1000 3888/3888 [==============================] - 1s 176us/sample - loss: 1.0210e-05 - val_loss: 7.5419e-06 Epoch 444/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 9.4297e-06 - val_loss: 3.3546e-06 Epoch 445/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 8.4666e-06 - val_loss: 1.2456e-05 Epoch 446/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0003e-05 - val_loss: 1.4192e-05 Epoch 447/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 6.1268e-06 - val_loss: 5.0171e-06 Epoch 448/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.0346e-05 - val_loss: 3.0288e-06 Epoch 449/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.0833e-05 - val_loss: 4.2388e-06 Epoch 450/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.7321e-06 - val_loss: 7.5831e-06 Epoch 451/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.6491e-06 - val_loss: 4.3831e-06 Epoch 452/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.3097e-05 - val_loss: 3.3071e-06 Epoch 453/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.0755e-06 - val_loss: 2.9045e-06 Epoch 454/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.7364e-06 - val_loss: 1.6519e-04 Epoch 455/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.6471e-05 - val_loss: 3.3185e-06 Epoch 456/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 6.3478e-06 - val_loss: 3.1858e-06 Epoch 457/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.5151e-05 - val_loss: 6.1860e-06 Epoch 458/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.1909e-06 - val_loss: 2.5943e-06 Epoch 459/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.5814e-06 - val_loss: 2.5300e-06 Epoch 460/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.0121e-06 - val_loss: 5.2197e-06 Epoch 461/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 7.6334e-06 - val_loss: 3.1797e-06 Epoch 462/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 1.2220e-05 - val_loss: 2.4944e-06 Epoch 463/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 6.5070e-06 - val_loss: 5.0679e-06 Epoch 464/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.0023e-06 - val_loss: 3.6424e-06 Epoch 465/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 8.2103e-06 - val_loss: 7.8455e-06 Epoch 466/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.2903e-05 - val_loss: 4.2873e-06 Epoch 467/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.9283e-06 - val_loss: 6.8694e-06 Epoch 468/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.3989e-05 - val_loss: 2.4157e-06 Epoch 469/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.9605e-06 - val_loss: 2.3979e-06 Epoch 470/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 4.5422e-06 - val_loss: 2.5331e-06 Epoch 471/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0753e-05 - val_loss: 4.2875e-06 Epoch 472/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.1282e-06 - val_loss: 5.3128e-06 Epoch 473/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.8463e-05 - val_loss: 5.5651e-06 Epoch 474/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.6825e-06 - val_loss: 3.9530e-06 Epoch 475/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.4487e-06 - val_loss: 2.8871e-06 Epoch 476/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.1087e-06 - val_loss: 3.5271e-06 Epoch 477/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.6146e-06 - val_loss: 3.2352e-06 Epoch 478/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 1.3658e-05 - val_loss: 5.3646e-06 Epoch 479/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.7163e-06 - val_loss: 3.0681e-06 Epoch 480/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.5382e-06 - val_loss: 4.4004e-06 Epoch 481/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.8956e-05 - val_loss: 5.7961e-06 Epoch 482/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 3.9107e-06 - val_loss: 8.8000e-06 Epoch 483/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 5.4816e-06 - val_loss: 3.2461e-06 Epoch 484/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 6.2144e-06 - val_loss: 5.8745e-06 Epoch 485/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.0546e-05 - val_loss: 2.2743e-06 Epoch 486/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.6837e-06 - val_loss: 4.7397e-06 Epoch 487/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.5805e-06 - val_loss: 3.7137e-06 Epoch 488/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.3241e-06 - val_loss: 6.6537e-05 Epoch 489/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.0601e-05 - val_loss: 2.2468e-06 Epoch 490/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.4153e-06 - val_loss: 3.3736e-06 Epoch 491/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 1.5456e-05 - val_loss: 3.6638e-06 Epoch 492/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.8104e-06 - val_loss: 9.0092e-06 Epoch 493/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 6.0428e-06 - val_loss: 5.4377e-06 Epoch 494/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.4788e-05 - val_loss: 2.6950e-06 Epoch 495/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 3.6352e-06 - val_loss: 2.4965e-06 Epoch 496/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 6.6340e-06 - val_loss: 2.2906e-06 Epoch 497/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 8.1215e-06 - val_loss: 4.0459e-06 Epoch 498/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.0471e-05 - val_loss: 3.3839e-06 Epoch 499/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 7.0745e-06 - val_loss: 3.0207e-06 Epoch 500/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.8430e-06 - val_loss: 2.7745e-06 Epoch 501/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 9.8741e-06 - val_loss: 3.5460e-06 Epoch 502/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.6537e-06 - val_loss: 2.8081e-06 Epoch 503/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.0974e-05 - val_loss: 3.2872e-06 Epoch 504/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.8955e-06 - val_loss: 8.9933e-06 Epoch 505/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.4767e-05 - val_loss: 4.4001e-06 Epoch 506/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.8573e-06 - val_loss: 2.9806e-06 Epoch 507/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 7.0499e-06 - val_loss: 3.7876e-06 Epoch 508/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 6.3418e-06 - val_loss: 2.6856e-06 Epoch 509/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 9.7561e-06 - val_loss: 9.4345e-05 Epoch 510/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.4799e-05 - val_loss: 2.6547e-06 Epoch 511/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.0276e-05 - val_loss: 2.6792e-06 Epoch 512/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 3.6023e-06 - val_loss: 3.2308e-06 Epoch 513/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 4.8425e-06 - val_loss: 1.8339e-05 Epoch 514/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 3.5220e-05 - val_loss: 6.7517e-06 Epoch 515/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.7801e-06 - val_loss: 2.6363e-06 Epoch 516/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.3591e-06 - val_loss: 2.6514e-06 Epoch 517/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 4.0640e-06 - val_loss: 2.2862e-06 Epoch 518/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.3175e-06 - val_loss: 2.9472e-06 Epoch 519/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 1.5639e-05 - val_loss: 2.5745e-06 Epoch 520/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.8680e-06 - val_loss: 2.6642e-06 Epoch 521/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.1189e-06 - val_loss: 3.6216e-06 Epoch 522/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 9.7338e-06 - val_loss: 2.3382e-06 Epoch 523/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 9.5729e-06 - val_loss: 5.4297e-06 Epoch 524/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.7782e-06 - val_loss: 2.9875e-06 Epoch 525/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 5.0834e-06 - val_loss: 4.2689e-06 Epoch 526/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 9.5475e-06 - val_loss: 4.4690e-05 Epoch 527/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.0988e-06 - val_loss: 9.9161e-06 Epoch 528/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 9.9685e-06 - val_loss: 4.0761e-06 Epoch 529/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.7720e-06 - val_loss: 4.3785e-06 Epoch 530/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.1965e-05 - val_loss: 2.0524e-05 Epoch 531/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 6.9731e-06 - val_loss: 5.0076e-04 Epoch 532/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 1.7071e-05 - val_loss: 5.1372e-06 Epoch 533/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.8901e-06 - val_loss: 1.1619e-05 Epoch 534/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.0359e-05 - val_loss: 8.2168e-06 Epoch 535/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.8009e-06 - val_loss: 2.0494e-06 Epoch 536/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.4112e-06 - val_loss: 1.7928e-05 Epoch 537/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 6.3608e-06 - val_loss: 5.3772e-06 Epoch 538/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.1524e-05 - val_loss: 1.6460e-05 Epoch 539/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 9.1070e-06 - val_loss: 2.4266e-06 Epoch 540/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.7735e-06 - val_loss: 3.1909e-06 Epoch 541/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.9909e-06 - val_loss: 2.9661e-04 Epoch 542/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.6526e-05 - val_loss: 3.1603e-06 Epoch 543/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.8250e-06 - val_loss: 9.4982e-06 Epoch 544/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 9.7077e-06 - val_loss: 6.6249e-06 Epoch 545/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.6057e-06 - val_loss: 1.6924e-05 Epoch 546/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.9536e-06 - val_loss: 8.7292e-06 Epoch 547/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.2411e-05 - val_loss: 9.0507e-06 Epoch 548/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 4.7158e-06 - val_loss: 3.6979e-06 Epoch 549/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 7.8583e-06 - val_loss: 3.5711e-05 Epoch 550/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 6.6038e-06 - val_loss: 2.7543e-06 Epoch 551/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 3.8270e-06 - val_loss: 3.0933e-05 Epoch 552/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.3328e-05 - val_loss: 6.3377e-06 Epoch 553/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.7365e-06 - val_loss: 3.5911e-06 Epoch 554/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.4481e-06 - val_loss: 2.3295e-05 Epoch 555/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.9009e-05 - val_loss: 5.3621e-06 Epoch 556/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.6942e-06 - val_loss: 4.2823e-06 Epoch 557/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.6321e-06 - val_loss: 2.9906e-06 Epoch 558/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.9592e-06 - val_loss: 3.2649e-06 Epoch 559/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.1233e-05 - val_loss: 3.9317e-06 Epoch 560/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.3850e-06 - val_loss: 4.1759e-06 Epoch 561/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.1517e-06 - val_loss: 2.4860e-06 Epoch 562/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 8.5320e-06 - val_loss: 3.1471e-06 Epoch 563/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 6.0377e-06 - val_loss: 4.1571e-06 Epoch 564/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 5.1724e-06 - val_loss: 2.2032e-06 Epoch 565/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.7054e-05 - val_loss: 7.4169e-06 Epoch 566/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.4974e-06 - val_loss: 2.2363e-06 Epoch 567/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.7101e-06 - val_loss: 4.5144e-06 Epoch 568/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.7443e-06 - val_loss: 2.3940e-06 Epoch 569/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.8161e-05 - val_loss: 2.7057e-06 Epoch 570/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 7.3577e-06 - val_loss: 2.8099e-06 Epoch 571/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 2.8868e-06 - val_loss: 5.6265e-06 Epoch 572/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.4986e-05 - val_loss: 2.0325e-06 Epoch 573/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 2.6073e-06 - val_loss: 2.9666e-06 Epoch 574/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 4.8887e-06 - val_loss: 4.5774e-05 Epoch 575/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 6.3607e-06 - val_loss: 4.2190e-06 Epoch 576/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 4.3793e-06 - val_loss: 4.3023e-06 Epoch 577/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 5.6023e-06 - val_loss: 2.6898e-06 Epoch 578/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 4.3601e-05 - val_loss: 2.1711e-06 Epoch 579/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.2702e-06 - val_loss: 3.4386e-06 Epoch 580/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.3803e-06 - val_loss: 2.6804e-06 Epoch 581/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 3.6102e-06 - val_loss: 2.6324e-06 Epoch 582/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.1063e-06 - val_loss: 3.0626e-06 Epoch 583/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 5.8641e-06 - val_loss: 3.8579e-06 Epoch 584/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 8.3281e-06 - val_loss: 2.6185e-06 Epoch 585/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 3.3884e-06 - val_loss: 3.7735e-06 Epoch 586/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 9.5669e-06 - val_loss: 2.8033e-06 Epoch 587/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.6300e-06 - val_loss: 7.1490e-06 Epoch 588/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.1629e-06 - val_loss: 3.3123e-06 Epoch 589/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.3849e-06 - val_loss: 2.2863e-05 Epoch 590/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.9293e-06 - val_loss: 2.5432e-06 Epoch 591/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.6427e-06 - val_loss: 4.2926e-06 Epoch 592/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 6.9737e-06 - val_loss: 1.6732e-05 Epoch 593/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 7.0126e-06 - val_loss: 1.6030e-05 Epoch 594/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.4753e-06 - val_loss: 2.9083e-06 Epoch 595/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.3914e-06 - val_loss: 9.0456e-06 Epoch 596/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 8.4843e-06 - val_loss: 9.1891e-05 Epoch 597/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.2993e-05 - val_loss: 2.5238e-06 Epoch 598/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.0747e-06 - val_loss: 3.6334e-06 Epoch 599/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.8808e-06 - val_loss: 9.1319e-05 Epoch 600/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.0723e-05 - val_loss: 2.4317e-06 Epoch 601/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.2033e-06 - val_loss: 4.2531e-06 Epoch 602/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.1684e-05 - val_loss: 1.7976e-05 Epoch 603/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 4.9644e-06 - val_loss: 2.3510e-06 Epoch 604/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.7155e-05 - val_loss: 3.3066e-06 Epoch 605/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.2901e-06 - val_loss: 4.9084e-06 Epoch 606/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 6.9689e-06 - val_loss: 2.2229e-06 Epoch 607/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 4.4699e-06 - val_loss: 9.7948e-06 Epoch 608/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.4524e-06 - val_loss: 1.2243e-05 Epoch 609/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.5679e-06 - val_loss: 7.9777e-06 Epoch 610/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.0225e-05 - val_loss: 2.5243e-06 Epoch 611/1000 3888/3888 [==============================] - 1s 159us/sample - loss: 5.9230e-06 - val_loss: 1.2367e-05 Epoch 612/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.1973e-06 - val_loss: 5.2115e-06 Epoch 613/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.0981e-05 - val_loss: 4.1662e-04 Epoch 614/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.0291e-05 - val_loss: 2.3225e-06 Epoch 615/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 2.5623e-06 - val_loss: 2.4138e-06 Epoch 616/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.4597e-06 - val_loss: 2.4034e-05 Epoch 617/1000 3888/3888 [==============================] - 1s 158us/sample - loss: 1.0837e-05 - val_loss: 2.4497e-06 Epoch 618/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.4676e-06 - val_loss: 5.3235e-06 Epoch 619/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.6024e-06 - val_loss: 3.0106e-06 Epoch 620/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 8.5510e-06 - val_loss: 2.0851e-06 Epoch 621/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.5331e-06 - val_loss: 7.1334e-06 Epoch 622/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.1532e-05 - val_loss: 1.0257e-05 Epoch 623/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.6361e-06 - val_loss: 2.2319e-06 Epoch 624/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.8293e-06 - val_loss: 2.1060e-06 Epoch 625/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.2566e-06 - val_loss: 2.8273e-06 Epoch 626/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.3481e-05 - val_loss: 2.8198e-06 Epoch 627/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.4382e-06 - val_loss: 2.7181e-06 Epoch 628/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 6.7636e-06 - val_loss: 2.9671e-06 Epoch 629/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.8254e-06 - val_loss: 2.9675e-06 Epoch 630/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.7969e-06 - val_loss: 2.3639e-06 Epoch 631/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 7.2364e-06 - val_loss: 2.4838e-06 Epoch 632/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.7791e-06 - val_loss: 3.1207e-06 Epoch 633/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.9815e-06 - val_loss: 2.3000e-06 Epoch 634/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 5.4541e-06 - val_loss: 2.4982e-06 Epoch 635/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.8936e-06 - val_loss: 2.0878e-06 Epoch 636/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 4.5526e-06 - val_loss: 2.1262e-06 Epoch 637/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 8.9842e-06 - val_loss: 2.7301e-06 Epoch 638/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 6.2820e-06 - val_loss: 7.1352e-06 Epoch 639/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.5106e-05 - val_loss: 3.7374e-05 Epoch 640/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.4692e-06 - val_loss: 2.5835e-06 Epoch 641/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.8777e-06 - val_loss: 1.9721e-06 Epoch 642/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 3.8143e-06 - val_loss: 3.7580e-06 Epoch 643/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.4087e-06 - val_loss: 1.8463e-06 Epoch 644/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 1.7205e-05 - val_loss: 2.2203e-06 Epoch 645/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.6050e-06 - val_loss: 2.2152e-06 Epoch 646/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.3518e-06 - val_loss: 8.0691e-06 Epoch 647/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.8656e-05 - val_loss: 2.5127e-06 Epoch 648/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.1149e-06 - val_loss: 3.2720e-06 Epoch 649/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.4947e-06 - val_loss: 3.1246e-06 Epoch 650/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 5.4091e-06 - val_loss: 1.9420e-06 Epoch 651/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.9869e-06 - val_loss: 6.0130e-06 Epoch 652/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.2578e-06 - val_loss: 4.0553e-05 Epoch 653/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 6.5127e-06 - val_loss: 2.4521e-06 Epoch 654/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.8958e-06 - val_loss: 1.9295e-06 Epoch 655/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 1.1211e-05 - val_loss: 2.9413e-05 Epoch 656/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.5038e-06 - val_loss: 2.9672e-06 Epoch 657/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 6.7257e-06 - val_loss: 2.7937e-06 Epoch 658/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 9.6235e-06 - val_loss: 6.0334e-06 Epoch 659/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 4.2697e-06 - val_loss: 2.7478e-06 Epoch 660/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.2434e-05 - val_loss: 2.0820e-05 Epoch 661/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 4.3218e-06 - val_loss: 4.3094e-05 Epoch 662/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.8367e-06 - val_loss: 7.1007e-06 Epoch 663/1000 3888/3888 [==============================] - 1s 153us/sample - loss: 8.0818e-06 - val_loss: 6.6116e-06 Epoch 664/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.2446e-06 - val_loss: 4.3292e-06 Epoch 665/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 2.7070e-06 - val_loss: 1.0022e-05 Epoch 666/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 8.0687e-06 - val_loss: 2.3745e-05 Epoch 667/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.3892e-05 - val_loss: 5.5405e-06 Epoch 668/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.9379e-06 - val_loss: 7.2324e-06 Epoch 669/1000 3888/3888 [==============================] - 1s 175us/sample - loss: 7.1673e-06 - val_loss: 2.9749e-06 Epoch 670/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 4.2551e-06 - val_loss: 3.6768e-06 Epoch 671/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 4.5692e-06 - val_loss: 1.4716e-05 Epoch 672/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 9.2295e-06 - val_loss: 2.6240e-06 Epoch 673/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.2107e-05 - val_loss: 1.8759e-06 Epoch 674/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 9.5258e-06 - val_loss: 5.5561e-05 Epoch 675/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.0844e-06 - val_loss: 1.9868e-06 Epoch 676/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.4193e-06 - val_loss: 5.3140e-06 Epoch 677/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.5394e-06 - val_loss: 2.0907e-06 Epoch 678/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.0830e-06 - val_loss: 1.9128e-05 Epoch 679/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.0754e-05 - val_loss: 1.8111e-06 Epoch 680/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.6430e-06 - val_loss: 4.8077e-06 Epoch 681/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.2270e-05 - val_loss: 3.3499e-06 Epoch 682/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.4881e-06 - val_loss: 3.5932e-06 Epoch 683/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 8.7650e-06 - val_loss: 8.9997e-06 Epoch 684/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.2323e-06 - val_loss: 2.5675e-06 Epoch 685/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.3130e-05 - val_loss: 5.0054e-06 Epoch 686/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.1298e-06 - val_loss: 1.7983e-06 Epoch 687/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.2279e-06 - val_loss: 2.6224e-06 Epoch 688/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.0000e-05 - val_loss: 1.1273e-05 Epoch 689/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.7483e-06 - val_loss: 2.3760e-06 Epoch 690/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.0569e-06 - val_loss: 1.9603e-04 Epoch 691/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.3105e-06 - val_loss: 1.9406e-06 Epoch 692/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.6690e-06 - val_loss: 4.9213e-06 Epoch 693/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 6.4126e-06 - val_loss: 5.5727e-06 Epoch 694/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.7468e-06 - val_loss: 2.6840e-06 Epoch 695/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 6.7808e-06 - val_loss: 1.6310e-04 Epoch 696/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.1829e-05 - val_loss: 3.1340e-06 Epoch 697/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.4261e-06 - val_loss: 2.6237e-06 Epoch 698/1000 3888/3888 [==============================] - 1s 160us/sample - loss: 3.5702e-06 - val_loss: 1.9178e-06 Epoch 699/1000 3888/3888 [==============================] - 1s 160us/sample - loss: 7.4502e-06 - val_loss: 3.7225e-06 Epoch 700/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 7.0083e-06 - val_loss: 1.2410e-05 Epoch 701/1000 3888/3888 [==============================] - 1s 160us/sample - loss: 1.1793e-05 - val_loss: 3.3129e-06 Epoch 702/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.6136e-06 - val_loss: 6.9046e-06 Epoch 703/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.1738e-05 - val_loss: 8.1944e-06 Epoch 704/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.2993e-06 - val_loss: 4.3426e-06 Epoch 705/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 4.0264e-06 - val_loss: 2.9858e-06 Epoch 706/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 5.5883e-06 - val_loss: 3.0930e-06 Epoch 707/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.6384e-06 - val_loss: 1.9893e-05 Epoch 708/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.2308e-05 - val_loss: 2.2723e-06 Epoch 709/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.4917e-06 - val_loss: 4.6097e-06 Epoch 710/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.5434e-06 - val_loss: 2.2377e-06 Epoch 711/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.5757e-06 - val_loss: 2.1981e-06 Epoch 712/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 5.5690e-06 - val_loss: 1.9117e-06 Epoch 713/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 7.6084e-06 - val_loss: 3.1844e-06 Epoch 714/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.9737e-06 - val_loss: 2.4693e-06 Epoch 715/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.2299e-06 - val_loss: 1.7391e-05 Epoch 716/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.4171e-06 - val_loss: 1.0237e-05 Epoch 717/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.0039e-05 - val_loss: 8.3671e-05 Epoch 718/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.0087e-05 - val_loss: 2.2911e-06 Epoch 719/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.3686e-06 - val_loss: 2.0400e-06 Epoch 720/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 7.8005e-06 - val_loss: 2.4595e-06 Epoch 721/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.7221e-06 - val_loss: 2.6916e-06 Epoch 722/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 9.2297e-06 - val_loss: 5.7757e-06 Epoch 723/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 3.6983e-06 - val_loss: 2.5092e-06 Epoch 724/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 9.0859e-06 - val_loss: 1.3905e-05 Epoch 725/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 6.1067e-06 - val_loss: 2.8300e-06 Epoch 726/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.2672e-06 - val_loss: 3.4278e-06 Epoch 727/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.6965e-06 - val_loss: 3.1145e-06 Epoch 728/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.5177e-05 - val_loss: 2.2529e-06 Epoch 729/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 7.2252e-06 - val_loss: 2.1555e-06 Epoch 730/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.8634e-06 - val_loss: 1.8502e-05 Epoch 731/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.6109e-06 - val_loss: 1.6665e-06 Epoch 732/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.4749e-06 - val_loss: 3.4787e-06 Epoch 733/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.6011e-06 - val_loss: 7.2570e-06 Epoch 734/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.7702e-06 - val_loss: 5.3229e-06 Epoch 735/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 1.2463e-05 - val_loss: 2.3051e-06 Epoch 736/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.8293e-06 - val_loss: 1.5986e-06 Epoch 737/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.1986e-06 - val_loss: 8.7274e-06 Epoch 738/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.5860e-06 - val_loss: 2.3562e-06 Epoch 739/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.8723e-06 - val_loss: 3.6557e-06 Epoch 740/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 8.5597e-06 - val_loss: 3.2005e-06 Epoch 741/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.6909e-06 - val_loss: 1.3349e-05 Epoch 742/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 1.7086e-05 - val_loss: 2.2086e-06 Epoch 743/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.2284e-06 - val_loss: 5.4729e-06 Epoch 744/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.8830e-06 - val_loss: 3.4757e-06 Epoch 745/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 6.9488e-06 - val_loss: 4.6278e-06 Epoch 746/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.2556e-06 - val_loss: 5.0094e-06 Epoch 747/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 6.9013e-06 - val_loss: 2.0116e-06 Epoch 748/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.1696e-05 - val_loss: 3.2718e-06 Epoch 749/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.7895e-06 - val_loss: 1.7937e-06 Epoch 750/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.6886e-06 - val_loss: 3.2168e-06 Epoch 751/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.0437e-06 - val_loss: 5.7697e-06 Epoch 752/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 6.8203e-06 - val_loss: 2.0313e-06 Epoch 753/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.9564e-06 - val_loss: 6.3141e-06 Epoch 754/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 6.0875e-06 - val_loss: 2.8463e-05 Epoch 755/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 1.8737e-05 - val_loss: 2.9726e-06 Epoch 756/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.1037e-06 - val_loss: 4.8715e-06 Epoch 757/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 2.2620e-06 - val_loss: 3.6488e-06 Epoch 758/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 4.2694e-06 - val_loss: 3.5380e-06 Epoch 759/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 1.0423e-05 - val_loss: 2.3417e-06 Epoch 760/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 7.2690e-06 - val_loss: 7.6126e-06 Epoch 761/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 5.1283e-06 - val_loss: 2.3281e-06 Epoch 762/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 3.7424e-06 - val_loss: 1.9560e-06 Epoch 763/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 5.0716e-06 - val_loss: 5.1792e-06 Epoch 764/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 1.0411e-05 - val_loss: 3.2560e-06 Epoch 765/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 6.6007e-06 - val_loss: 1.2112e-05 Epoch 766/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 6.1528e-06 - val_loss: 2.1560e-06 Epoch 767/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.7834e-06 - val_loss: 2.7505e-06 Epoch 768/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 6.0295e-06 - val_loss: 3.5703e-06 Epoch 769/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 9.8851e-06 - val_loss: 4.5103e-06 Epoch 770/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.6137e-06 - val_loss: 5.0688e-05 Epoch 771/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.9347e-06 - val_loss: 1.8578e-06 Epoch 772/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.8994e-06 - val_loss: 3.4737e-06 Epoch 773/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.1411e-06 - val_loss: 2.3996e-06 Epoch 774/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 7.5358e-06 - val_loss: 9.5618e-06 Epoch 775/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.3155e-06 - val_loss: 5.0994e-06 Epoch 776/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.4584e-06 - val_loss: 2.0672e-06 Epoch 777/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.9611e-06 - val_loss: 4.6562e-06 Epoch 778/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.5168e-06 - val_loss: 1.4829e-06 Epoch 779/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 9.2897e-06 - val_loss: 4.8001e-06 Epoch 780/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.2998e-06 - val_loss: 4.3309e-06 Epoch 781/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.5847e-06 - val_loss: 8.4400e-06 Epoch 782/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.3959e-06 - val_loss: 5.8241e-05 Epoch 783/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 7.6059e-06 - val_loss: 2.2832e-06 Epoch 784/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.5673e-06 - val_loss: 2.8397e-06 Epoch 785/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.5470e-05 - val_loss: 2.6250e-06 Epoch 786/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.1511e-06 - val_loss: 3.4366e-06 Epoch 787/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.1752e-06 - val_loss: 2.7688e-06 Epoch 788/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 5.3467e-06 - val_loss: 6.1376e-06 Epoch 789/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.3430e-06 - val_loss: 1.8958e-06 Epoch 790/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.3303e-06 - val_loss: 5.4627e-06 Epoch 791/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 9.2902e-06 - val_loss: 7.9086e-06 Epoch 792/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 9.4993e-06 - val_loss: 4.4534e-06 Epoch 793/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.2254e-06 - val_loss: 8.9466e-06 Epoch 794/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.7849e-06 - val_loss: 6.6420e-06 Epoch 795/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 7.6654e-06 - val_loss: 3.2498e-06 Epoch 796/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 7.6201e-06 - val_loss: 1.8296e-05 Epoch 797/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 7.4328e-06 - val_loss: 6.3781e-06 Epoch 798/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 5.7137e-06 - val_loss: 2.0336e-06 Epoch 799/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 3.0551e-06 - val_loss: 1.3844e-05 Epoch 800/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.0619e-05 - val_loss: 1.7140e-05 Epoch 801/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 5.1146e-06 - val_loss: 7.6266e-06 Epoch 802/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.6091e-06 - val_loss: 4.6041e-06 Epoch 803/1000 3888/3888 [==============================] - 1s 160us/sample - loss: 9.5688e-06 - val_loss: 2.4609e-06 Epoch 804/1000 3888/3888 [==============================] - 1s 160us/sample - loss: 3.0091e-06 - val_loss: 2.1043e-06 Epoch 805/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 6.2905e-06 - val_loss: 2.7834e-06 Epoch 806/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 6.3295e-06 - val_loss: 1.5077e-05 Epoch 807/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.7134e-06 - val_loss: 2.2854e-06 Epoch 808/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 1.4392e-05 - val_loss: 3.7948e-06 Epoch 809/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 2.4486e-06 - val_loss: 9.5667e-06 Epoch 810/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.2050e-06 - val_loss: 9.7170e-06 Epoch 811/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 7.2433e-06 - val_loss: 2.9135e-06 Epoch 812/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.1999e-06 - val_loss: 3.7421e-06 Epoch 813/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.1655e-05 - val_loss: 2.4722e-06 Epoch 814/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.4282e-06 - val_loss: 5.6160e-06 Epoch 815/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.3381e-06 - val_loss: 3.8852e-06 Epoch 816/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 4.6563e-06 - val_loss: 9.2089e-06 Epoch 817/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 7.7846e-06 - val_loss: 2.0207e-06 Epoch 818/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 2.5294e-06 - val_loss: 3.9383e-06 Epoch 819/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.2620e-06 - val_loss: 2.9572e-06 Epoch 820/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 1.7502e-05 - val_loss: 1.9467e-06 Epoch 821/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.1084e-06 - val_loss: 4.7428e-06 Epoch 822/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.5549e-06 - val_loss: 9.1445e-06 Epoch 823/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.5881e-06 - val_loss: 2.3637e-06 Epoch 824/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.9334e-06 - val_loss: 3.1392e-06 Epoch 825/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 9.5863e-06 - val_loss: 2.0007e-06 Epoch 826/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.2421e-05 - val_loss: 2.0864e-06 Epoch 827/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.1206e-06 - val_loss: 2.4137e-06 Epoch 828/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.5770e-06 - val_loss: 4.1399e-06 Epoch 829/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0052e-05 - val_loss: 1.1299e-04 Epoch 830/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.7765e-06 - val_loss: 1.8016e-06 Epoch 831/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.0182e-06 - val_loss: 2.2805e-06 Epoch 832/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 3.0929e-06 - val_loss: 2.1839e-06 Epoch 833/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 6.7647e-06 - val_loss: 7.4611e-06 Epoch 834/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.8294e-06 - val_loss: 2.1110e-06 Epoch 835/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 8.6005e-06 - val_loss: 2.5510e-06 Epoch 836/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 2.1009e-06 - val_loss: 4.8268e-06 Epoch 837/1000 3888/3888 [==============================] - 1s 173us/sample - loss: 7.8965e-06 - val_loss: 6.7347e-06 Epoch 838/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.3969e-06 - val_loss: 2.3511e-06 Epoch 839/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.0199e-05 - val_loss: 2.6520e-06 Epoch 840/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.7535e-06 - val_loss: 2.9943e-06 Epoch 841/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.3540e-06 - val_loss: 2.8025e-06 Epoch 842/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.0953e-05 - val_loss: 1.4128e-06 Epoch 843/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.1547e-06 - val_loss: 2.5967e-06 Epoch 844/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.1284e-06 - val_loss: 4.5697e-06 Epoch 845/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 6.2030e-06 - val_loss: 2.9638e-06 Epoch 846/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 6.7587e-06 - val_loss: 2.0322e-06 Epoch 847/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.2278e-06 - val_loss: 5.1796e-06 Epoch 848/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.1164e-05 - val_loss: 1.9499e-06 Epoch 849/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.7049e-06 - val_loss: 2.8449e-06 Epoch 850/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 1.6122e-05 - val_loss: 3.3790e-05 Epoch 851/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.0325e-06 - val_loss: 1.4015e-06 Epoch 852/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 4.6529e-06 - val_loss: 1.4511e-06 Epoch 853/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 2.2267e-06 - val_loss: 3.1475e-06 Epoch 854/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.6639e-06 - val_loss: 3.1564e-05 Epoch 855/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 7.0384e-06 - val_loss: 3.1758e-06 Epoch 856/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 1.0654e-05 - val_loss: 2.9913e-06 Epoch 857/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.4661e-06 - val_loss: 1.3029e-06 Epoch 858/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.8052e-06 - val_loss: 7.6139e-06 Epoch 859/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.5722e-06 - val_loss: 2.8367e-05 Epoch 860/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.5544e-06 - val_loss: 1.2088e-05 Epoch 861/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 5.1299e-06 - val_loss: 2.9148e-06 Epoch 862/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 9.3464e-06 - val_loss: 3.2254e-05 Epoch 863/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.1837e-06 - val_loss: 1.4637e-06 Epoch 864/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 2.1099e-06 - val_loss: 2.0721e-06 Epoch 865/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 7.5095e-06 - val_loss: 3.8327e-06 Epoch 866/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 3.6482e-06 - val_loss: 7.2715e-06 Epoch 867/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.6010e-06 - val_loss: 5.7788e-05 Epoch 868/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 6.3916e-06 - val_loss: 1.0005e-05 Epoch 869/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 6.2830e-06 - val_loss: 1.6308e-06 Epoch 870/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.3658e-06 - val_loss: 2.4649e-06 Epoch 871/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 8.7231e-06 - val_loss: 7.7415e-06 Epoch 872/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.8396e-06 - val_loss: 8.2775e-06 Epoch 873/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.1010e-06 - val_loss: 8.1543e-06 Epoch 874/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.6043e-06 - val_loss: 1.7351e-06 Epoch 875/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 8.4081e-06 - val_loss: 3.2645e-06 Epoch 876/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 4.0041e-06 - val_loss: 2.1227e-06 Epoch 877/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 1.2496e-05 - val_loss: 3.9363e-06 Epoch 878/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.1125e-06 - val_loss: 1.4032e-05 Epoch 879/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.7346e-06 - val_loss: 4.6809e-06 Epoch 880/1000 3888/3888 [==============================] - 1s 161us/sample - loss: 6.8622e-06 - val_loss: 4.6001e-06 Epoch 881/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.8242e-06 - val_loss: 2.2288e-06 Epoch 882/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.3127e-06 - val_loss: 1.9237e-06 Epoch 883/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 4.0922e-06 - val_loss: 1.7614e-06 Epoch 884/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 5.8579e-06 - val_loss: 4.6058e-06 Epoch 885/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 9.8404e-06 - val_loss: 1.3702e-06 Epoch 886/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 1.9641e-06 - val_loss: 3.1501e-06 Epoch 887/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 4.7577e-06 - val_loss: 2.7389e-06 Epoch 888/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.4902e-05 - val_loss: 7.0653e-06 Epoch 889/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 2.0851e-06 - val_loss: 1.7295e-06 Epoch 890/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.7044e-06 - val_loss: 1.9539e-06 Epoch 891/1000 3888/3888 [==============================] - 1s 158us/sample - loss: 3.0548e-06 - val_loss: 1.7713e-06 Epoch 892/1000 3888/3888 [==============================] - 1s 160us/sample - loss: 2.5777e-06 - val_loss: 3.0858e-06 Epoch 893/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 9.8964e-06 - val_loss: 4.3673e-06 Epoch 894/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.4679e-06 - val_loss: 1.4053e-06 Epoch 895/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.3370e-06 - val_loss: 2.3683e-06 Epoch 896/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 6.9820e-06 - val_loss: 2.4671e-06 Epoch 897/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.2697e-06 - val_loss: 3.8837e-06 Epoch 898/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 7.1294e-06 - val_loss: 1.5857e-05 Epoch 899/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.7637e-06 - val_loss: 2.5314e-05 Epoch 900/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 4.0619e-06 - val_loss: 3.0389e-06 Epoch 901/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 5.5887e-06 - val_loss: 5.8972e-06 Epoch 902/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 7.2578e-06 - val_loss: 1.8121e-06 Epoch 903/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 2.5340e-05 - val_loss: 1.9057e-06 Epoch 904/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.6592e-06 - val_loss: 1.3450e-06 Epoch 905/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.3561e-06 - val_loss: 1.6019e-06 Epoch 906/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.5925e-06 - val_loss: 4.7183e-06 Epoch 907/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 4.8295e-06 - val_loss: 1.0839e-05 Epoch 908/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 2.7352e-06 - val_loss: 3.6938e-06 Epoch 909/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 3.7493e-06 - val_loss: 5.1690e-06 Epoch 910/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 5.9894e-06 - val_loss: 2.5425e-05 Epoch 911/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.4853e-05 - val_loss: 4.6569e-06 Epoch 912/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.5471e-06 - val_loss: 3.3032e-06 Epoch 913/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 3.1720e-06 - val_loss: 3.1734e-06 Epoch 914/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 3.1823e-06 - val_loss: 4.0763e-06 Epoch 915/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.6357e-05 - val_loss: 1.5139e-06 Epoch 916/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.4681e-06 - val_loss: 1.5220e-06 Epoch 917/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.9307e-06 - val_loss: 1.3779e-06 Epoch 918/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 1.5191e-05 - val_loss: 3.8788e-06 Epoch 919/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 1.7677e-06 - val_loss: 4.1557e-06 Epoch 920/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 3.5336e-06 - val_loss: 1.4581e-05 Epoch 921/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.7531e-06 - val_loss: 8.8573e-06 Epoch 922/1000 3888/3888 [==============================] - 1s 177us/sample - loss: 3.7063e-06 - val_loss: 5.4229e-06 Epoch 923/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.0532e-05 - val_loss: 1.7654e-06 Epoch 924/1000 3888/3888 [==============================] - 1s 174us/sample - loss: 5.5829e-06 - val_loss: 1.9428e-06 Epoch 925/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 2.7866e-06 - val_loss: 9.8100e-06 Epoch 926/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 1.0816e-05 - val_loss: 1.7568e-06 Epoch 927/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 2.0562e-06 - val_loss: 4.1915e-06 Epoch 928/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 5.6562e-06 - val_loss: 2.1544e-06 Epoch 929/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.9427e-06 - val_loss: 4.1118e-06 Epoch 930/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 5.1345e-06 - val_loss: 3.0541e-06 Epoch 931/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 5.2787e-06 - val_loss: 1.5460e-05 Epoch 932/1000 3888/3888 [==============================] - 1s 171us/sample - loss: 1.6622e-05 - val_loss: 9.7273e-06 Epoch 933/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 2.4480e-06 - val_loss: 1.6005e-06 Epoch 934/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.3727e-06 - val_loss: 1.7770e-06 Epoch 935/1000 3888/3888 [==============================] - 1s 164us/sample - loss: 3.6466e-06 - val_loss: 2.2452e-06 Epoch 936/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.9288e-06 - val_loss: 1.4226e-06 Epoch 937/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 4.4455e-06 - val_loss: 2.3434e-06 Epoch 938/1000 3888/3888 [==============================] - 1s 167us/sample - loss: 7.6965e-06 - val_loss: 3.0733e-06 Epoch 939/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 3.0304e-06 - val_loss: 5.1687e-06 Epoch 940/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 6.8483e-06 - val_loss: 1.7941e-06 Epoch 941/1000 3888/3888 [==============================] - 1s 169us/sample - loss: 5.6291e-06 - val_loss: 4.0073e-06 Epoch 942/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 1.1260e-05 - val_loss: 6.6182e-06 Epoch 943/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.4625e-06 - val_loss: 1.5301e-06 Epoch 944/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.5897e-06 - val_loss: 2.4575e-06 Epoch 945/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 9.6700e-06 - val_loss: 3.8631e-06 Epoch 946/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.5730e-06 - val_loss: 5.1494e-06 Epoch 947/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.1154e-06 - val_loss: 5.0446e-06 Epoch 948/1000 3888/3888 [==============================] - 1s 168us/sample - loss: 7.4122e-06 - val_loss: 2.4344e-05 Epoch 949/1000 3888/3888 [==============================] - 1s 163us/sample - loss: 3.0407e-06 - val_loss: 1.8600e-06 Epoch 950/1000 3888/3888 [==============================] - 1s 165us/sample - loss: 4.1492e-06 - val_loss: 1.9007e-06 Epoch 951/1000 3888/3888 [==============================] - 1s 172us/sample - loss: 5.8114e-06 - val_loss: 1.1793e-05 Epoch 952/1000 3888/3888 [==============================] - 1s 170us/sample - loss: 3.5881e-06 - val_loss: 3.4288e-06 Epoch 953/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 5.6412e-06 - val_loss: 2.0950e-06 Epoch 954/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 7.5404e-06 - val_loss: 2.9750e-06 Epoch 955/1000 3888/3888 [==============================] - 1s 162us/sample - loss: 6.3202e-06 - val_loss: 1.8026e-05 Epoch 956/1000 3888/3888 [==============================] - 1s 166us/sample - loss: 2.5637e-06 - val_loss: 8.9821e-06 Epoch 957/1000 3616/3888 [==========================>...] - ETA: 0s - loss: 5.6745e-06Restoring model weights from the end of the best epoch. 3888/3888 [==============================] - 1s 166us/sample - loss: 6.0992e-06 - val_loss: 2.3747e-05 Epoch 00957: early stopping
In [46]:
print(history.history.keys())
print('best value: ', autoencoder.evaluate(X_train_1D_norm, X_train_1D_norm, verbose=0))
pd.DataFrame(history.history).plot(figsize=(8, 5), logy=True)
plt.grid()
dict_keys(['loss', 'val_loss']) best value: 1.3029401012924306e-06
In [47]:
X_reconstructions = autoencoder.predict(X_train_1D_norm)
X_reconstructions = stdscaler.inverse_transform(X_reconstructions)
In [48]:
calculateerror(X_train_1D.reshape(len(times),len(groups),nl,nc),
X_reconstructions.reshape(len(times),len(groups),nl,nc),
groups,
print_step=0)
max_abs_error: 11.947998046875 mean_abs_error: 0.018922157238844824
/home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: divide by zero encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until /home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until
In [49]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(X_reconstructions.reshape(len(times),len(groups),nl,nc)[100,i,:,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [50]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_train_1D[:,i*nl*nc+4])
ax.flatten()[i].plot(times, X_reconstructions[:,i*nl*nc+4],'--')
ax.flatten()[i].set_title(group)
PCA Non-linear autoencoder¶
In [51]:
from sklearn.decomposition import PCA
pca = PCA(n_components=115)
X_train_pca = pca.fit_transform(X_train_1D)
In [52]:
for i, s in enumerate(pca.singular_values_):
print(i,s)
0 80968.84766232983 1 18736.416256363147 2 13673.705906493908 3 5582.676911420286 4 3850.103044758365 5 1800.019343080367 6 1528.5522031979297 7 1280.5509296354314 8 732.9460285895258 9 435.26801371011277 10 297.1644606962294 11 181.60261273092226 12 87.67587246299661 13 64.75204695667632 14 56.14576255903517 15 22.805565355182708 16 14.680854372754887 17 9.433904365360851 18 8.086013100546879 19 6.691068654905932 20 3.580681950463192 21 3.20088686807251 22 2.9496893087417306 23 1.3593700244681008 24 1.303570351076243 25 0.9841955380229419 26 0.8745552427561293 27 0.6751953659028838 28 0.45335331999123346 29 0.3682248610588608 30 0.34029883857464993 31 0.30108410171505134 32 0.2747720601623404 33 0.2580691302875648 34 0.25155019818685037 35 0.23908841104002337 36 0.231200546275467 37 0.2264119861037694 38 0.2231818070561315 39 0.22132948289422547 40 0.21754946426500482 41 0.1964041659361343 42 0.1679801080275365 43 0.15181483405154386 44 0.14977231392571658 45 0.1473383360201773 46 0.14617716679239093 47 0.1453544112672034 48 0.14453011030302515 49 0.1439177002926129 50 0.14160316373151513 51 0.13906581761656092 52 0.13780621050453318 53 0.13740573893534178 54 0.13666581093000651 55 0.13493310745883458 56 0.1332568496983295 57 0.13291836640430776 58 0.1307904036435818 59 0.13070402948183749 60 0.1292305310358126 61 0.1279535385999949 62 0.12723542380683203 63 0.1268304817673247 64 0.12614436133224236 65 0.12489907480215162 66 0.12360679751365647 67 0.1219607164200381 68 0.12142235784823262 69 0.11993204400816027 70 0.11937139251378552 71 0.11747907576501383 72 0.11617850864659925 73 0.1155368010649835 74 0.1144785725500825 75 0.1128507232536807 76 0.11206714316404931 77 0.11071054748169312 78 0.10711683369304364 79 0.10560161003677218 80 0.0007051765903159103 81 0.0007049224859099373 82 0.00011454367898059194 83 9.983968383730085e-05 84 9.968887287803789e-05 85 7.40362947509492e-05 86 7.08901089558992e-05 87 7.08383287927755e-05 88 7.07760356261245e-05 89 7.074115206083078e-05 90 7.046377936715971e-05 91 7.03625872480673e-05 92 7.00733821003764e-05 93 4.467147195296282e-05 94 2.8083779709223102e-05 95 7.047996459632438e-06 96 5.655737100821686e-06 97 5.947717937151221e-07 98 4.883996219172943e-07 99 2.3762879983255165e-07 100 2.1838297456611626e-07 101 1.9162186180281583e-07 102 1.644840238306863e-07 103 1.4385138440225086e-07 104 1.352797419456935e-07 105 1.232880930271776e-07 106 1.1190936811411953e-07 107 1.025810141391199e-07 108 9.480668974466976e-08 109 8.381246249793427e-08 110 7.394636004087645e-08 111 4.5573931030267163e-08 112 3.9864671094623444e-08 113 1.501245627371161e-08 114 8.090413105687467e-12
In [53]:
np.random.seed(42)
tf.random.set_seed(42)
# Need to have validation loss
early_stopping = keras.callbacks.EarlyStopping(monitor='val_loss',
min_delta=0.0,
patience=100,
verbose=2,
restore_best_weights=True)
encoder = keras.models.Sequential([keras.layers.Dense(100, input_shape=[115], activation="elu"),
keras.layers.Dense(50, activation="elu"),
keras.layers.Dense(15)])
decoder = keras.models.Sequential([keras.layers.Dense(50, input_shape=[15], activation="elu"),
keras.layers.Dense(100, activation="elu"),
keras.layers.Dense(115),
])
autoencoder = keras.models.Sequential([encoder, decoder])
autoencoder.compile(loss="mse",
optimizer=keras.optimizers.Nadam(lr=0.0003, beta_1=0.9, beta_2=0.999)
)
encoder.summary()
decoder.summary()
Model: "sequential_9" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= dense_12 (Dense) (None, 100) 11600 _________________________________________________________________ dense_13 (Dense) (None, 50) 5050 _________________________________________________________________ dense_14 (Dense) (None, 15) 765 ================================================================= Total params: 17,415 Trainable params: 17,415 Non-trainable params: 0 _________________________________________________________________ Model: "sequential_10" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= dense_15 (Dense) (None, 50) 800 _________________________________________________________________ dense_16 (Dense) (None, 100) 5100 _________________________________________________________________ dense_17 (Dense) (None, 115) 11615 ================================================================= Total params: 17,515 Trainable params: 17,515 Non-trainable params: 0 _________________________________________________________________
In [54]:
history = autoencoder.fit(X_train_pca,
X_train_pca,
epochs=1000,
validation_data=(X_train_pca, X_train_pca),
callbacks=[early_stopping])
Train on 3888 samples, validate on 3888 samples Epoch 1/1000 3888/3888 [==============================] - 1s 350us/sample - loss: 8216.8399 - val_loss: 1035.5376 Epoch 2/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 734.8472 - val_loss: 478.6631 Epoch 3/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 287.6702 - val_loss: 164.8730 Epoch 4/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 114.3151 - val_loss: 85.6140 Epoch 5/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 74.4078 - val_loss: 64.5016 Epoch 6/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 56.8069 - val_loss: 48.2749 Epoch 7/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 39.9972 - val_loss: 31.4465 Epoch 8/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 27.3115 - val_loss: 26.8397 Epoch 9/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 24.1810 - val_loss: 20.3460 Epoch 10/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 20.6134 - val_loss: 20.0034 Epoch 11/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 18.6640 - val_loss: 16.0814 Epoch 12/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 17.6694 - val_loss: 15.7350 Epoch 13/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 17.3316 - val_loss: 14.4572 Epoch 14/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 15.0136 - val_loss: 19.0059 Epoch 15/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 15.0771 - val_loss: 22.0533 Epoch 16/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 18.5594 - val_loss: 12.7963 Epoch 17/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 12.9352 - val_loss: 18.6959 Epoch 18/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 13.1217 - val_loss: 12.4854 Epoch 19/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 12.0637 - val_loss: 12.3078 Epoch 20/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 12.3319 - val_loss: 11.1737 Epoch 21/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 10.7176 - val_loss: 11.1456 Epoch 22/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 11.7356 - val_loss: 10.3615 Epoch 23/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 11.0722 - val_loss: 18.2247 Epoch 24/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 9.6776 - val_loss: 8.0109 Epoch 25/1000 3888/3888 [==============================] - 0s 104us/sample - loss: 10.0381 - val_loss: 9.3924 Epoch 26/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 8.5657 - val_loss: 7.8695 Epoch 27/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 8.5826 - val_loss: 7.0682 Epoch 28/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 6.6980 - val_loss: 10.2599 Epoch 29/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 7.9348 - val_loss: 6.7623 Epoch 30/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 6.3146 - val_loss: 11.6546 Epoch 31/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 6.6887 - val_loss: 4.7955 Epoch 32/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 6.2727 - val_loss: 6.3317 Epoch 33/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 5.0163 - val_loss: 4.6039 Epoch 34/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 5.0681 - val_loss: 4.4893 Epoch 35/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 6.2797 - val_loss: 5.4832 Epoch 36/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 4.2529 - val_loss: 7.7428 Epoch 37/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 5.7423 - val_loss: 3.6946 Epoch 38/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 3.7593 - val_loss: 4.4625 Epoch 39/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 6.1392 - val_loss: 3.2314 Epoch 40/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 3.1810 - val_loss: 5.2926 Epoch 41/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 3.9926 - val_loss: 3.0844 Epoch 42/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 4.7274 - val_loss: 3.1449 Epoch 43/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 3.1673 - val_loss: 3.3040 Epoch 44/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 3.6463 - val_loss: 6.1856 Epoch 45/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 3.1306 - val_loss: 2.8394 Epoch 46/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 3.6238 - val_loss: 3.8606 Epoch 47/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 3.5124 - val_loss: 2.4359 Epoch 48/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 3.1312 - val_loss: 3.6617 Epoch 49/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 6.2187 - val_loss: 2.4693 Epoch 50/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 2.3745 - val_loss: 2.2090 Epoch 51/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 2.5612 - val_loss: 2.2610 Epoch 52/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 3.2424 - val_loss: 2.4355 Epoch 53/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 3.2387 - val_loss: 3.8992 Epoch 54/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 2.1877 - val_loss: 2.1899 Epoch 55/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 3.2998 - val_loss: 2.1290 Epoch 56/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 2.2326 - val_loss: 3.1886 Epoch 57/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 2.5787 - val_loss: 1.8565 Epoch 58/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 2.1722 - val_loss: 3.6203 Epoch 59/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 2.7243 - val_loss: 2.6876 Epoch 60/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 2.7190 - val_loss: 3.9519 Epoch 61/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 2.0679 - val_loss: 2.4134 Epoch 62/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 2.8736 - val_loss: 2.9162 Epoch 63/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 2.3045 - val_loss: 1.5566 Epoch 64/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 1.8647 - val_loss: 1.8459 Epoch 65/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 1.8960 - val_loss: 1.6027 Epoch 66/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 2.2757 - val_loss: 2.0049 Epoch 67/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 2.5586 - val_loss: 1.8275 Epoch 68/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.5149 - val_loss: 3.1557 Epoch 69/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.9558 - val_loss: 5.4694 Epoch 70/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 2.0542 - val_loss: 1.3017 Epoch 71/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 2.0808 - val_loss: 1.3416 Epoch 72/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.7672 - val_loss: 6.4220 Epoch 73/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 2.0566 - val_loss: 1.5075 Epoch 74/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 1.3479 - val_loss: 1.2734 Epoch 75/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 2.8215 - val_loss: 1.1642 Epoch 76/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 1.2874 - val_loss: 1.6600 Epoch 77/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.6195 - val_loss: 1.8074 Epoch 78/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.4400 - val_loss: 1.1657 Epoch 79/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.7366 - val_loss: 1.2417 Epoch 80/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 1.6294 - val_loss: 1.1398 Epoch 81/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.3111 - val_loss: 1.8351 Epoch 82/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 1.6109 - val_loss: 1.2803 Epoch 83/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 1.5509 - val_loss: 1.2131 Epoch 84/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.5350 - val_loss: 1.2273 Epoch 85/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.8093 - val_loss: 1.1503 Epoch 86/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 1.4707 - val_loss: 0.8966 Epoch 87/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 1.8914 - val_loss: 1.4310 Epoch 88/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 1.0821 - val_loss: 1.0744 Epoch 89/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.5161 - val_loss: 4.8746 Epoch 90/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.4543 - val_loss: 1.4334 Epoch 91/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.4222 - val_loss: 1.4939 Epoch 92/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 1.1731 - val_loss: 1.4988 Epoch 93/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 1.1756 - val_loss: 0.9266 Epoch 94/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.4469 - val_loss: 1.0253 Epoch 95/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.8176 - val_loss: 0.8664 Epoch 96/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.9648 - val_loss: 1.0377 Epoch 97/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.1351 - val_loss: 3.4181 Epoch 98/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 1.3361 - val_loss: 1.0682 Epoch 99/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 1.4501 - val_loss: 1.7614 Epoch 100/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 1.0410 - val_loss: 1.0233 Epoch 101/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 1.0815 - val_loss: 1.1423 Epoch 102/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 1.3637 - val_loss: 2.3318 Epoch 103/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 1.1442 - val_loss: 1.0397 Epoch 104/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 1.2897 - val_loss: 0.8084 Epoch 105/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.0296 - val_loss: 0.8900 Epoch 106/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.6222 - val_loss: 1.1878 Epoch 107/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.1859 - val_loss: 1.0418 Epoch 108/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.8976 - val_loss: 0.9111 Epoch 109/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.8155 - val_loss: 0.8472 Epoch 110/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 0.7887 - val_loss: 0.8219 Epoch 111/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.9838 - val_loss: 1.0034 Epoch 112/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.9982 - val_loss: 1.6208 Epoch 113/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.9630 - val_loss: 1.6363 Epoch 114/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 1.1791 - val_loss: 0.9612 Epoch 115/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.9612 - val_loss: 2.0543 Epoch 116/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.9296 - val_loss: 0.9522 Epoch 117/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 1.3498 - val_loss: 1.1120 Epoch 118/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.0312 - val_loss: 0.7799 Epoch 119/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.9143 - val_loss: 0.8475 Epoch 120/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.9657 - val_loss: 0.8492 Epoch 121/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.0142 - val_loss: 2.9638 Epoch 122/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 1.0105 - val_loss: 1.1665 Epoch 123/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.2598 - val_loss: 0.6654 Epoch 124/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.3521 - val_loss: 4.6589 Epoch 125/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.7538 - val_loss: 0.6329 Epoch 126/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.6902 - val_loss: 1.1616 Epoch 127/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.9577 - val_loss: 1.2226 Epoch 128/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.8392 - val_loss: 1.5217 Epoch 129/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 1.0894 - val_loss: 0.7025 Epoch 130/1000 3888/3888 [==============================] - 0s 105us/sample - loss: 0.7085 - val_loss: 3.8074 Epoch 131/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 1.0549 - val_loss: 1.5405 Epoch 132/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.9562 - val_loss: 0.6383 Epoch 133/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.6945 - val_loss: 1.1428 Epoch 134/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.9887 - val_loss: 0.9811 Epoch 135/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.8502 - val_loss: 1.8200 Epoch 136/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 1.1426 - val_loss: 0.5464 Epoch 137/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.7358 - val_loss: 2.9150 Epoch 138/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.8490 - val_loss: 0.5434 Epoch 139/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.6872 - val_loss: 0.6051 Epoch 140/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.9126 - val_loss: 0.8126 Epoch 141/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.7404 - val_loss: 0.5384 Epoch 142/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 1.1845 - val_loss: 0.7142 Epoch 143/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.6414 - val_loss: 0.6885 Epoch 144/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.7735 - val_loss: 0.5251 Epoch 145/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.7328 - val_loss: 0.6404 Epoch 146/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.9080 - val_loss: 0.9105 Epoch 147/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.7505 - val_loss: 0.6369 Epoch 148/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.7557 - val_loss: 0.8681 Epoch 149/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.6928 - val_loss: 1.3653 Epoch 150/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.6620 - val_loss: 2.3189 Epoch 151/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.0340 - val_loss: 0.9976 Epoch 152/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.9673 - val_loss: 0.4628 Epoch 153/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.6260 - val_loss: 0.5743 Epoch 154/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.6380 - val_loss: 0.5684 Epoch 155/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.6946 - val_loss: 1.1423 Epoch 156/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.7326 - val_loss: 0.4652 Epoch 157/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 1.0057 - val_loss: 6.7324 Epoch 158/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.6059 - val_loss: 0.6790 Epoch 159/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.8157 - val_loss: 0.6086 Epoch 160/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5720 - val_loss: 0.7312 Epoch 161/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 1.0858 - val_loss: 0.7631 Epoch 162/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.5493 - val_loss: 0.5134 Epoch 163/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5508 - val_loss: 0.4698 Epoch 164/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5643 - val_loss: 1.5089 Epoch 165/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.6657 - val_loss: 0.7396 Epoch 166/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 1.0131 - val_loss: 0.7039 Epoch 167/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.5875 - val_loss: 0.8609 Epoch 168/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.6684 - val_loss: 0.4834 Epoch 169/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.6034 - val_loss: 0.4738 Epoch 170/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5876 - val_loss: 0.5982 Epoch 171/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.6162 - val_loss: 0.4592 Epoch 172/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.9735 - val_loss: 3.9072 Epoch 173/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.5683 - val_loss: 0.7324 Epoch 174/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4736 - val_loss: 0.6436 Epoch 175/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.7635 - val_loss: 0.5221 Epoch 176/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.5022 - val_loss: 0.5284 Epoch 177/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.9081 - val_loss: 0.3940 Epoch 178/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.5417 - val_loss: 0.6338 Epoch 179/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.6509 - val_loss: 0.5283 Epoch 180/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5980 - val_loss: 0.4690 Epoch 181/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.4802 - val_loss: 1.9992 Epoch 182/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.6410 - val_loss: 0.4436 Epoch 183/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.8637 - val_loss: 4.1249 Epoch 184/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5879 - val_loss: 0.3693 Epoch 185/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.4946 - val_loss: 0.4360 Epoch 186/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4871 - val_loss: 0.3611 Epoch 187/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.8429 - val_loss: 0.7500 Epoch 188/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4741 - val_loss: 0.4966 Epoch 189/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.5152 - val_loss: 0.4412 Epoch 190/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.5302 - val_loss: 0.5422 Epoch 191/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.6159 - val_loss: 0.7860 Epoch 192/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.9343 - val_loss: 0.3714 Epoch 193/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3797 - val_loss: 0.4119 Epoch 194/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.9614 - val_loss: 1.1046 Epoch 195/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4618 - val_loss: 0.3230 Epoch 196/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.3828 - val_loss: 0.7817 Epoch 197/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.6175 - val_loss: 0.3456 Epoch 198/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3847 - val_loss: 0.3844 Epoch 199/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5813 - val_loss: 0.3922 Epoch 200/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.6021 - val_loss: 0.3374 Epoch 201/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.4812 - val_loss: 0.4325 Epoch 202/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.6186 - val_loss: 0.3100 Epoch 203/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.5373 - val_loss: 0.3897 Epoch 204/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.6631 - val_loss: 0.3726 Epoch 205/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3617 - val_loss: 1.1166 Epoch 206/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.5764 - val_loss: 0.3329 Epoch 207/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4529 - val_loss: 0.4618 Epoch 208/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.5060 - val_loss: 1.1717 Epoch 209/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.6154 - val_loss: 0.3748 Epoch 210/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5711 - val_loss: 0.3233 Epoch 211/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4372 - val_loss: 1.0693 Epoch 212/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.5397 - val_loss: 0.4055 Epoch 213/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.5550 - val_loss: 0.4316 Epoch 214/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.4883 - val_loss: 1.7524 Epoch 215/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.5401 - val_loss: 0.2791 Epoch 216/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3796 - val_loss: 0.3850 Epoch 217/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5851 - val_loss: 0.7781 Epoch 218/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4598 - val_loss: 0.3279 Epoch 219/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4044 - val_loss: 0.8433 Epoch 220/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5715 - val_loss: 0.5616 Epoch 221/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.5648 - val_loss: 0.3714 Epoch 222/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5455 - val_loss: 0.4319 Epoch 223/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3983 - val_loss: 0.5357 Epoch 224/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.6957 - val_loss: 0.2656 Epoch 225/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3856 - val_loss: 0.3817 Epoch 226/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.4755 - val_loss: 0.4138 Epoch 227/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4430 - val_loss: 0.2712 Epoch 228/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4744 - val_loss: 0.4675 Epoch 229/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.4947 - val_loss: 0.2937 Epoch 230/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4393 - val_loss: 0.2935 Epoch 231/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.7524 - val_loss: 1.0370 Epoch 232/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3340 - val_loss: 0.8246 Epoch 233/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5298 - val_loss: 0.4243 Epoch 234/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3120 - val_loss: 0.2703 Epoch 235/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4062 - val_loss: 0.6819 Epoch 236/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5868 - val_loss: 0.4592 Epoch 237/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4618 - val_loss: 0.5651 Epoch 238/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.4403 - val_loss: 0.3998 Epoch 239/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4485 - val_loss: 0.2443 Epoch 240/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.5178 - val_loss: 0.3245 Epoch 241/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3858 - val_loss: 0.2397 Epoch 242/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5087 - val_loss: 0.4250 Epoch 243/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.3187 - val_loss: 0.6091 Epoch 244/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4811 - val_loss: 0.3518 Epoch 245/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4983 - val_loss: 0.3290 Epoch 246/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4482 - val_loss: 13.6777 Epoch 247/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.6413 - val_loss: 0.2412 Epoch 248/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.4357 - val_loss: 0.7732 Epoch 249/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.3026 - val_loss: 0.2616 Epoch 250/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5362 - val_loss: 0.3769 Epoch 251/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2905 - val_loss: 0.2573 Epoch 252/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3270 - val_loss: 0.3514 Epoch 253/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.4122 - val_loss: 0.2422 Epoch 254/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4569 - val_loss: 0.3905 Epoch 255/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4234 - val_loss: 0.8740 Epoch 256/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3737 - val_loss: 0.3015 Epoch 257/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4645 - val_loss: 0.2305 Epoch 258/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.4733 - val_loss: 0.2199 Epoch 259/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.5157 - val_loss: 0.3957 Epoch 260/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3820 - val_loss: 0.2170 Epoch 261/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.4262 - val_loss: 0.6148 Epoch 262/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3734 - val_loss: 0.5266 Epoch 263/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3270 - val_loss: 0.2189 Epoch 264/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3403 - val_loss: 0.2379 Epoch 265/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4678 - val_loss: 0.3071 Epoch 266/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.5544 - val_loss: 0.3708 Epoch 267/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3362 - val_loss: 0.4895 Epoch 268/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.4220 - val_loss: 0.3320 Epoch 269/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3395 - val_loss: 3.4882 Epoch 270/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.7336 - val_loss: 0.1818 Epoch 271/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.2746 - val_loss: 0.2618 Epoch 272/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3936 - val_loss: 0.3034 Epoch 273/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4184 - val_loss: 0.2212 Epoch 274/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3998 - val_loss: 0.4902 Epoch 275/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3726 - val_loss: 0.4604 Epoch 276/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4985 - val_loss: 0.2128 Epoch 277/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2766 - val_loss: 0.2204 Epoch 278/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3875 - val_loss: 0.3277 Epoch 279/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3580 - val_loss: 0.2699 Epoch 280/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.4392 - val_loss: 0.1791 Epoch 281/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3801 - val_loss: 0.1798 Epoch 282/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.5200 - val_loss: 0.3491 Epoch 283/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.4499 - val_loss: 0.1996 Epoch 284/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2098 - val_loss: 0.3724 Epoch 285/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2777 - val_loss: 0.2067 Epoch 286/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5630 - val_loss: 0.4450 Epoch 287/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3493 - val_loss: 1.4603 Epoch 288/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3166 - val_loss: 1.2091 Epoch 289/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3779 - val_loss: 0.2556 Epoch 290/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.4501 - val_loss: 0.2847 Epoch 291/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3842 - val_loss: 0.3794 Epoch 292/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3835 - val_loss: 0.1885 Epoch 293/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.5568 - val_loss: 0.2272 Epoch 294/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2366 - val_loss: 0.3628 Epoch 295/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5458 - val_loss: 0.2561 Epoch 296/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3127 - val_loss: 0.1622 Epoch 297/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2576 - val_loss: 0.3122 Epoch 298/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2810 - val_loss: 0.2176 Epoch 299/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2867 - val_loss: 1.3051 Epoch 300/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5521 - val_loss: 0.2558 Epoch 301/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.6692 - val_loss: 2.4660 Epoch 302/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3079 - val_loss: 0.1847 Epoch 303/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3328 - val_loss: 0.3259 Epoch 304/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2076 - val_loss: 0.5820 Epoch 305/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3806 - val_loss: 0.2453 Epoch 306/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3200 - val_loss: 0.9052 Epoch 307/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.4105 - val_loss: 0.2222 Epoch 308/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3260 - val_loss: 0.2016 Epoch 309/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4173 - val_loss: 0.2092 Epoch 310/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2555 - val_loss: 0.2004 Epoch 311/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4442 - val_loss: 0.8916 Epoch 312/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3581 - val_loss: 0.3374 Epoch 313/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3195 - val_loss: 0.1720 Epoch 314/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5692 - val_loss: 0.1761 Epoch 315/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2894 - val_loss: 0.9929 Epoch 316/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2409 - val_loss: 0.2485 Epoch 317/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2499 - val_loss: 1.7151 Epoch 318/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4223 - val_loss: 0.1690 Epoch 319/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4611 - val_loss: 0.4686 Epoch 320/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3102 - val_loss: 0.6225 Epoch 321/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3214 - val_loss: 0.6102 Epoch 322/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3736 - val_loss: 0.2135 Epoch 323/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2744 - val_loss: 0.5275 Epoch 324/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3305 - val_loss: 0.5360 Epoch 325/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.4953 - val_loss: 1.5719 Epoch 326/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2914 - val_loss: 0.2915 Epoch 327/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.5582 - val_loss: 0.2534 Epoch 328/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1686 - val_loss: 0.1849 Epoch 329/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3183 - val_loss: 0.2378 Epoch 330/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2638 - val_loss: 0.2549 Epoch 331/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4683 - val_loss: 0.5738 Epoch 332/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2462 - val_loss: 0.3669 Epoch 333/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3551 - val_loss: 0.3539 Epoch 334/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3676 - val_loss: 0.2149 Epoch 335/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3822 - val_loss: 0.2677 Epoch 336/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3148 - val_loss: 0.1874 Epoch 337/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3052 - val_loss: 0.1729 Epoch 338/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3294 - val_loss: 0.5296 Epoch 339/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2549 - val_loss: 0.1566 Epoch 340/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.6803 - val_loss: 0.1528 Epoch 341/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2820 - val_loss: 0.1621 Epoch 342/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2569 - val_loss: 0.2948 Epoch 343/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2777 - val_loss: 0.1793 Epoch 344/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2643 - val_loss: 0.1675 Epoch 345/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 1.0457 - val_loss: 0.2280 Epoch 346/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.1634 - val_loss: 0.2097 Epoch 347/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.2629 - val_loss: 0.1523 Epoch 348/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.2267 - val_loss: 0.3109 Epoch 349/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.3566 - val_loss: 0.5713 Epoch 350/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4100 - val_loss: 0.3455 Epoch 351/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2356 - val_loss: 0.2165 Epoch 352/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3050 - val_loss: 0.2068 Epoch 353/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3370 - val_loss: 0.1490 Epoch 354/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2280 - val_loss: 0.2326 Epoch 355/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2450 - val_loss: 0.4718 Epoch 356/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.5532 - val_loss: 0.1571 Epoch 357/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4088 - val_loss: 0.1935 Epoch 358/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2749 - val_loss: 0.3091 Epoch 359/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2141 - val_loss: 2.6155 Epoch 360/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.4179 - val_loss: 0.1446 Epoch 361/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1693 - val_loss: 0.2741 Epoch 362/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3763 - val_loss: 0.2877 Epoch 363/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2834 - val_loss: 0.4913 Epoch 364/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4395 - val_loss: 0.1369 Epoch 365/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1879 - val_loss: 0.3937 Epoch 366/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5547 - val_loss: 0.4815 Epoch 367/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2271 - val_loss: 0.1674 Epoch 368/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.6336 - val_loss: 0.2553 Epoch 369/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2331 - val_loss: 0.1477 Epoch 370/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1887 - val_loss: 0.2659 Epoch 371/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.1769 - val_loss: 0.2451 Epoch 372/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.4905 - val_loss: 0.1483 Epoch 373/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2491 - val_loss: 1.4189 Epoch 374/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3997 - val_loss: 0.5958 Epoch 375/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.1781 - val_loss: 0.3190 Epoch 376/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3223 - val_loss: 0.3323 Epoch 377/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3008 - val_loss: 0.2113 Epoch 378/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2688 - val_loss: 0.1849 Epoch 379/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2543 - val_loss: 0.3723 Epoch 380/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3643 - val_loss: 0.4242 Epoch 381/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.6252 - val_loss: 0.2200 Epoch 382/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2029 - val_loss: 0.1415 Epoch 383/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3383 - val_loss: 0.1782 Epoch 384/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3775 - val_loss: 0.1580 Epoch 385/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1713 - val_loss: 0.2144 Epoch 386/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3011 - val_loss: 0.1277 Epoch 387/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3260 - val_loss: 0.1441 Epoch 388/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.5931 - val_loss: 0.1462 Epoch 389/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2501 - val_loss: 0.1594 Epoch 390/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2001 - val_loss: 0.6000 Epoch 391/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3451 - val_loss: 0.1337 Epoch 392/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.4132 - val_loss: 0.7836 Epoch 393/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1698 - val_loss: 0.2283 Epoch 394/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2597 - val_loss: 0.1958 Epoch 395/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2388 - val_loss: 0.5857 Epoch 396/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4306 - val_loss: 0.4839 Epoch 397/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2447 - val_loss: 0.2404 Epoch 398/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4029 - val_loss: 0.1449 Epoch 399/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2828 - val_loss: 0.1557 Epoch 400/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.1864 - val_loss: 0.3945 Epoch 401/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.4236 - val_loss: 0.8957 Epoch 402/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.2442 - val_loss: 0.2266 Epoch 403/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2858 - val_loss: 1.3159 Epoch 404/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 0.2758 - val_loss: 0.1848 Epoch 405/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2350 - val_loss: 0.1491 Epoch 406/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.7253 - val_loss: 0.3110 Epoch 407/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.1595 - val_loss: 0.1586 Epoch 408/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2119 - val_loss: 0.1952 Epoch 409/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2796 - val_loss: 0.2310 Epoch 410/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.4064 - val_loss: 0.1475 Epoch 411/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2106 - val_loss: 0.1843 Epoch 412/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2952 - val_loss: 0.1461 Epoch 413/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3154 - val_loss: 0.1183 Epoch 414/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2349 - val_loss: 0.5241 Epoch 415/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2721 - val_loss: 0.2537 Epoch 416/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.4350 - val_loss: 0.2941 Epoch 417/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2948 - val_loss: 0.1705 Epoch 418/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3082 - val_loss: 0.1806 Epoch 419/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1497 - val_loss: 0.1812 Epoch 420/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3611 - val_loss: 0.1523 Epoch 421/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.5096 - val_loss: 0.3445 Epoch 422/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1854 - val_loss: 0.1631 Epoch 423/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3808 - val_loss: 0.8352 Epoch 424/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2614 - val_loss: 0.1819 Epoch 425/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1556 - val_loss: 0.2947 Epoch 426/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2776 - val_loss: 0.1609 Epoch 427/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2473 - val_loss: 6.5086 Epoch 428/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.5162 - val_loss: 0.2375 Epoch 429/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2138 - val_loss: 0.1557 Epoch 430/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1518 - val_loss: 0.1666 Epoch 431/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.4123 - val_loss: 0.4097 Epoch 432/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2520 - val_loss: 0.1487 Epoch 433/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2125 - val_loss: 0.7565 Epoch 434/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3642 - val_loss: 0.1863 Epoch 435/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2973 - val_loss: 0.1553 Epoch 436/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3322 - val_loss: 0.1894 Epoch 437/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2354 - val_loss: 0.2549 Epoch 438/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3874 - val_loss: 0.7970 Epoch 439/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2733 - val_loss: 0.4693 Epoch 440/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2804 - val_loss: 0.1339 Epoch 441/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2992 - val_loss: 0.1933 Epoch 442/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3579 - val_loss: 0.6725 Epoch 443/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2456 - val_loss: 0.1213 Epoch 444/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.5065 - val_loss: 0.1683 Epoch 445/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1760 - val_loss: 0.7920 Epoch 446/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1805 - val_loss: 0.2318 Epoch 447/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2686 - val_loss: 1.0969 Epoch 448/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.7978 - val_loss: 0.1065 Epoch 449/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1744 - val_loss: 0.1161 Epoch 450/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1513 - val_loss: 0.1948 Epoch 451/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2213 - val_loss: 0.3021 Epoch 452/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2769 - val_loss: 0.7772 Epoch 453/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2520 - val_loss: 0.0964 Epoch 454/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1834 - val_loss: 0.2704 Epoch 455/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.5091 - val_loss: 0.2693 Epoch 456/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1670 - val_loss: 0.1048 Epoch 457/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2408 - val_loss: 0.4624 Epoch 458/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2231 - val_loss: 0.5321 Epoch 459/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3421 - val_loss: 0.1725 Epoch 460/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.6578 - val_loss: 0.1178 Epoch 461/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.1462 - val_loss: 0.1687 Epoch 462/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1669 - val_loss: 0.1088 Epoch 463/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2523 - val_loss: 0.2806 Epoch 464/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.1608 - val_loss: 0.1107 Epoch 465/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 0.3310 - val_loss: 0.1148 Epoch 466/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4867 - val_loss: 0.1165 Epoch 467/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.1935 - val_loss: 0.1175 Epoch 468/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1562 - val_loss: 0.1751 Epoch 469/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3961 - val_loss: 0.1083 Epoch 470/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1986 - val_loss: 0.1122 Epoch 471/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2485 - val_loss: 0.6166 Epoch 472/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4213 - val_loss: 0.1524 Epoch 473/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.2002 - val_loss: 0.1236 Epoch 474/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3508 - val_loss: 1.5711 Epoch 475/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.1837 - val_loss: 0.1309 Epoch 476/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2277 - val_loss: 0.1131 Epoch 477/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2975 - val_loss: 0.2854 Epoch 478/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2614 - val_loss: 0.5924 Epoch 479/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3192 - val_loss: 0.1016 Epoch 480/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.5707 - val_loss: 0.1934 Epoch 481/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.1189 - val_loss: 0.1644 Epoch 482/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2568 - val_loss: 0.1097 Epoch 483/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1757 - val_loss: 0.4679 Epoch 484/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2128 - val_loss: 0.1927 Epoch 485/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2827 - val_loss: 0.2093 Epoch 486/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3117 - val_loss: 0.1397 Epoch 487/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3040 - val_loss: 0.4850 Epoch 488/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2419 - val_loss: 0.0983 Epoch 489/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3133 - val_loss: 0.1047 Epoch 490/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2546 - val_loss: 0.1812 Epoch 491/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1543 - val_loss: 0.1438 Epoch 492/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3562 - val_loss: 0.5370 Epoch 493/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2411 - val_loss: 0.2459 Epoch 494/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2796 - val_loss: 0.1305 Epoch 495/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2355 - val_loss: 0.1168 Epoch 496/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1503 - val_loss: 0.3642 Epoch 497/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.4564 - val_loss: 0.1099 Epoch 498/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3604 - val_loss: 0.2479 Epoch 499/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1472 - val_loss: 0.0989 Epoch 500/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2040 - val_loss: 0.7074 Epoch 501/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2872 - val_loss: 0.2423 Epoch 502/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2268 - val_loss: 1.9836 Epoch 503/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3945 - val_loss: 0.1856 Epoch 504/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2157 - val_loss: 0.7675 Epoch 505/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2056 - val_loss: 0.1101 Epoch 506/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.5522 - val_loss: 0.1454 Epoch 507/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1237 - val_loss: 0.0859 Epoch 508/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1996 - val_loss: 0.1441 Epoch 509/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3209 - val_loss: 1.6737 Epoch 510/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3048 - val_loss: 0.2290 Epoch 511/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1731 - val_loss: 0.0884 Epoch 512/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3418 - val_loss: 0.2255 Epoch 513/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2397 - val_loss: 0.1540 Epoch 514/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.4559 - val_loss: 0.2381 Epoch 515/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1352 - val_loss: 0.1025 Epoch 516/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2467 - val_loss: 0.3032 Epoch 517/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2010 - val_loss: 0.1537 Epoch 518/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2950 - val_loss: 0.2778 Epoch 519/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2308 - val_loss: 0.1241 Epoch 520/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2325 - val_loss: 0.1332 Epoch 521/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2722 - val_loss: 0.3358 Epoch 522/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2279 - val_loss: 0.1489 Epoch 523/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3582 - val_loss: 0.1367 Epoch 524/1000 3888/3888 [==============================] - 0s 99us/sample - loss: 0.2329 - val_loss: 0.2969 Epoch 525/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2159 - val_loss: 0.3246 Epoch 526/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1950 - val_loss: 0.3147 Epoch 527/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3956 - val_loss: 7.3494 Epoch 528/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3437 - val_loss: 0.0881 Epoch 529/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1196 - val_loss: 0.1449 Epoch 530/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.4482 - val_loss: 0.1258 Epoch 531/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.1242 - val_loss: 0.6483 Epoch 532/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2776 - val_loss: 0.1437 Epoch 533/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2503 - val_loss: 1.2579 Epoch 534/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3776 - val_loss: 0.3604 Epoch 535/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.1576 - val_loss: 0.1602 Epoch 536/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1559 - val_loss: 1.1819 Epoch 537/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2492 - val_loss: 1.3400 Epoch 538/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2191 - val_loss: 0.2714 Epoch 539/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2219 - val_loss: 0.2076 Epoch 540/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.4030 - val_loss: 0.3182 Epoch 541/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3296 - val_loss: 0.2039 Epoch 542/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1752 - val_loss: 0.0873 Epoch 543/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1601 - val_loss: 0.1252 Epoch 544/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3413 - val_loss: 0.1242 Epoch 545/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3152 - val_loss: 0.5218 Epoch 546/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.4251 - val_loss: 0.0797 Epoch 547/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1196 - val_loss: 0.1614 Epoch 548/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1196 - val_loss: 0.1528 Epoch 549/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2530 - val_loss: 0.1248 Epoch 550/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1898 - val_loss: 0.1503 Epoch 551/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3436 - val_loss: 0.1489 Epoch 552/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.6686 - val_loss: 0.0900 Epoch 553/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1087 - val_loss: 0.0915 Epoch 554/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.2307 - val_loss: 0.1497 Epoch 555/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1315 - val_loss: 0.1125 Epoch 556/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2178 - val_loss: 0.0999 Epoch 557/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2228 - val_loss: 0.1494 Epoch 558/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2624 - val_loss: 0.1095 Epoch 559/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3272 - val_loss: 0.0776 Epoch 560/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1515 - val_loss: 0.0950 Epoch 561/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3040 - val_loss: 0.2647 Epoch 562/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1793 - val_loss: 0.1014 Epoch 563/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3359 - val_loss: 0.1184 Epoch 564/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1899 - val_loss: 0.0942 Epoch 565/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3715 - val_loss: 0.1087 Epoch 566/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2037 - val_loss: 0.0782 Epoch 567/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3112 - val_loss: 0.0978 Epoch 568/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2408 - val_loss: 0.3924 Epoch 569/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1590 - val_loss: 1.3941 Epoch 570/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2905 - val_loss: 0.1515 Epoch 571/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1788 - val_loss: 0.1330 Epoch 572/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.7504 - val_loss: 0.1255 Epoch 573/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.1324 - val_loss: 0.0807 Epoch 574/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3505 - val_loss: 0.1954 Epoch 575/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1591 - val_loss: 0.1588 Epoch 576/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1638 - val_loss: 0.1217 Epoch 577/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1684 - val_loss: 1.1948 Epoch 578/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3209 - val_loss: 2.4540 Epoch 579/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2027 - val_loss: 0.2902 Epoch 580/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1824 - val_loss: 1.1407 Epoch 581/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.7238 - val_loss: 0.7343 Epoch 582/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2528 - val_loss: 0.1349 Epoch 583/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.0870 - val_loss: 0.1277 Epoch 584/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1854 - val_loss: 0.0752 Epoch 585/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1600 - val_loss: 0.2389 Epoch 586/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2054 - val_loss: 0.2099 Epoch 587/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1946 - val_loss: 0.0964 Epoch 588/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1801 - val_loss: 0.1505 Epoch 589/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3357 - val_loss: 0.2940 Epoch 590/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1841 - val_loss: 0.0980 Epoch 591/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2601 - val_loss: 1.4146 Epoch 592/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2359 - val_loss: 0.2659 Epoch 593/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3359 - val_loss: 0.9762 Epoch 594/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2604 - val_loss: 0.2707 Epoch 595/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.1782 - val_loss: 0.0883 Epoch 596/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.1563 - val_loss: 0.3305 Epoch 597/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3133 - val_loss: 0.1402 Epoch 598/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2273 - val_loss: 0.0909 Epoch 599/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2835 - val_loss: 0.3548 Epoch 600/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1437 - val_loss: 0.1741 Epoch 601/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.5701 - val_loss: 0.0724 Epoch 602/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2331 - val_loss: 0.0757 Epoch 603/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1205 - val_loss: 0.1547 Epoch 604/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1852 - val_loss: 0.0996 Epoch 605/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1500 - val_loss: 1.2783 Epoch 606/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3526 - val_loss: 0.1634 Epoch 607/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2808 - val_loss: 0.2861 Epoch 608/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.1209 - val_loss: 0.1037 Epoch 609/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2524 - val_loss: 0.2760 Epoch 610/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.1978 - val_loss: 1.0580 Epoch 611/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3811 - val_loss: 0.1306 Epoch 612/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1923 - val_loss: 0.0874 Epoch 613/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1602 - val_loss: 2.0825 Epoch 614/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2804 - val_loss: 0.0988 Epoch 615/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1910 - val_loss: 0.1996 Epoch 616/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2529 - val_loss: 0.0950 Epoch 617/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2094 - val_loss: 0.1801 Epoch 618/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2606 - val_loss: 0.2317 Epoch 619/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1509 - val_loss: 0.3306 Epoch 620/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2635 - val_loss: 1.3363 Epoch 621/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2510 - val_loss: 0.1044 Epoch 622/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.2769 - val_loss: 0.0698 Epoch 623/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2682 - val_loss: 0.0758 Epoch 624/1000 3888/3888 [==============================] - 0s 122us/sample - loss: 0.2395 - val_loss: 0.0754 Epoch 625/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.1595 - val_loss: 0.0751 Epoch 626/1000 3888/3888 [==============================] - 0s 121us/sample - loss: 0.3488 - val_loss: 0.2506 Epoch 627/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2114 - val_loss: 0.3729 Epoch 628/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.3222 - val_loss: 0.0774 Epoch 629/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.1160 - val_loss: 0.0774 Epoch 630/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2245 - val_loss: 1.0920 Epoch 631/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1637 - val_loss: 0.7492 Epoch 632/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2871 - val_loss: 0.0790 Epoch 633/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2022 - val_loss: 0.0768 Epoch 634/1000 3888/3888 [==============================] - 0s 101us/sample - loss: 0.4095 - val_loss: 0.0741 Epoch 635/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.1436 - val_loss: 0.1450 Epoch 636/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2784 - val_loss: 0.1240 Epoch 637/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.3168 - val_loss: 0.1197 Epoch 638/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.1192 - val_loss: 0.1027 Epoch 639/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2754 - val_loss: 0.1046 Epoch 640/1000 3888/3888 [==============================] - 0s 108us/sample - loss: 0.2155 - val_loss: 0.1262 Epoch 641/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2931 - val_loss: 0.0998 Epoch 642/1000 3888/3888 [==============================] - 0s 106us/sample - loss: 0.2829 - val_loss: 0.2800 Epoch 643/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.1867 - val_loss: 0.0856 Epoch 644/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.1415 - val_loss: 0.0848 Epoch 645/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2062 - val_loss: 0.1192 Epoch 646/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2964 - val_loss: 0.9496 Epoch 647/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3719 - val_loss: 0.0943 Epoch 648/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2579 - val_loss: 0.0845 Epoch 649/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1479 - val_loss: 0.1502 Epoch 650/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2403 - val_loss: 0.0886 Epoch 651/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1478 - val_loss: 0.1243 Epoch 652/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1717 - val_loss: 1.0306 Epoch 653/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3565 - val_loss: 0.8566 Epoch 654/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1678 - val_loss: 0.1520 Epoch 655/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3637 - val_loss: 0.0870 Epoch 656/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.0890 - val_loss: 0.0814 Epoch 657/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2048 - val_loss: 0.8283 Epoch 658/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.4130 - val_loss: 0.0890 Epoch 659/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2712 - val_loss: 11.0017 Epoch 660/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2800 - val_loss: 0.0646 Epoch 661/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2036 - val_loss: 0.1236 Epoch 662/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1792 - val_loss: 0.3171 Epoch 663/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1702 - val_loss: 0.2309 Epoch 664/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1761 - val_loss: 0.0936 Epoch 665/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3059 - val_loss: 0.0795 Epoch 666/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1339 - val_loss: 0.5195 Epoch 667/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3182 - val_loss: 0.7800 Epoch 668/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1705 - val_loss: 0.1097 Epoch 669/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3370 - val_loss: 0.1058 Epoch 670/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1714 - val_loss: 0.0946 Epoch 671/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2334 - val_loss: 0.1269 Epoch 672/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1296 - val_loss: 0.6107 Epoch 673/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.3355 - val_loss: 1.4171 Epoch 674/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.2077 - val_loss: 0.0893 Epoch 675/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.3296 - val_loss: 0.0989 Epoch 676/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1053 - val_loss: 0.0674 Epoch 677/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1628 - val_loss: 0.0900 Epoch 678/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2998 - val_loss: 0.1535 Epoch 679/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2699 - val_loss: 0.0844 Epoch 680/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.1186 - val_loss: 0.1922 Epoch 681/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1664 - val_loss: 0.1743 Epoch 682/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2901 - val_loss: 0.3231 Epoch 683/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.2731 - val_loss: 0.3195 Epoch 684/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1579 - val_loss: 0.1066 Epoch 685/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2696 - val_loss: 0.5095 Epoch 686/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1651 - val_loss: 0.1047 Epoch 687/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2215 - val_loss: 0.3594 Epoch 688/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3110 - val_loss: 0.1079 Epoch 689/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1413 - val_loss: 0.0932 Epoch 690/1000 3888/3888 [==============================] - 0s 111us/sample - loss: 0.5393 - val_loss: 1.0025 Epoch 691/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1635 - val_loss: 0.0791 Epoch 692/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1216 - val_loss: 0.1672 Epoch 693/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1872 - val_loss: 1.9027 Epoch 694/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3784 - val_loss: 0.1756 Epoch 695/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.0907 - val_loss: 0.1522 Epoch 696/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3989 - val_loss: 0.4528 Epoch 697/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1232 - val_loss: 1.3107 Epoch 698/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1884 - val_loss: 0.0924 Epoch 699/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1757 - val_loss: 0.0892 Epoch 700/1000 3888/3888 [==============================] - 0s 109us/sample - loss: 0.2306 - val_loss: 0.8284 Epoch 701/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2986 - val_loss: 0.0940 Epoch 702/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.3809 - val_loss: 0.0645 Epoch 703/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1149 - val_loss: 0.1171 Epoch 704/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3491 - val_loss: 0.1278 Epoch 705/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0949 - val_loss: 0.1971 Epoch 706/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1540 - val_loss: 0.0819 Epoch 707/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1636 - val_loss: 0.1790 Epoch 708/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3759 - val_loss: 0.0761 Epoch 709/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1004 - val_loss: 0.5051 Epoch 710/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3742 - val_loss: 0.0927 Epoch 711/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0919 - val_loss: 0.0770 Epoch 712/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2373 - val_loss: 0.0684 Epoch 713/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2051 - val_loss: 0.2519 Epoch 714/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1275 - val_loss: 0.1482 Epoch 715/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2786 - val_loss: 0.0798 Epoch 716/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1782 - val_loss: 0.1037 Epoch 717/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2454 - val_loss: 0.3674 Epoch 718/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2237 - val_loss: 0.0922 Epoch 719/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2857 - val_loss: 2.2270 Epoch 720/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3365 - val_loss: 0.0757 Epoch 721/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1209 - val_loss: 0.2106 Epoch 722/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.2216 - val_loss: 0.4283 Epoch 723/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1889 - val_loss: 0.1084 Epoch 724/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1584 - val_loss: 0.3512 Epoch 725/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2541 - val_loss: 0.0756 Epoch 726/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2104 - val_loss: 1.1149 Epoch 727/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.4546 - val_loss: 0.1372 Epoch 728/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.0930 - val_loss: 0.1230 Epoch 729/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.2025 - val_loss: 0.2054 Epoch 730/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.1965 - val_loss: 0.1469 Epoch 731/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1679 - val_loss: 0.1350 Epoch 732/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.3183 - val_loss: 0.2085 Epoch 733/1000 3888/3888 [==============================] - 0s 107us/sample - loss: 0.1354 - val_loss: 0.1995 Epoch 734/1000 3888/3888 [==============================] - 0s 110us/sample - loss: 0.4262 - val_loss: 0.1180 Epoch 735/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.0822 - val_loss: 0.0714 Epoch 736/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.1160 - val_loss: 0.0657 Epoch 737/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3000 - val_loss: 2.2415 Epoch 738/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3901 - val_loss: 0.0762 Epoch 739/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.0749 - val_loss: 0.0842 Epoch 740/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.2134 - val_loss: 0.0962 Epoch 741/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2142 - val_loss: 0.0832 Epoch 742/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.3354 - val_loss: 0.0695 Epoch 743/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.1033 - val_loss: 0.0970 Epoch 744/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3971 - val_loss: 0.1142 Epoch 745/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1268 - val_loss: 0.0800 Epoch 746/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1142 - val_loss: 0.6437 Epoch 747/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1911 - val_loss: 0.0937 Epoch 748/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2800 - val_loss: 0.0614 Epoch 749/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.1228 - val_loss: 0.1064 Epoch 750/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3614 - val_loss: 0.0861 Epoch 751/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1369 - val_loss: 0.1685 Epoch 752/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.1699 - val_loss: 0.1814 Epoch 753/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2921 - val_loss: 0.4110 Epoch 754/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.1310 - val_loss: 0.2202 Epoch 755/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2092 - val_loss: 0.1008 Epoch 756/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.2689 - val_loss: 0.1771 Epoch 757/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2158 - val_loss: 0.0685 Epoch 758/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.1553 - val_loss: 0.6060 Epoch 759/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2235 - val_loss: 1.8171 Epoch 760/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2023 - val_loss: 0.2343 Epoch 761/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.3096 - val_loss: 0.0948 Epoch 762/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.1600 - val_loss: 0.0636 Epoch 763/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.1146 - val_loss: 0.3618 Epoch 764/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.3212 - val_loss: 0.0661 Epoch 765/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2220 - val_loss: 0.0773 Epoch 766/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.4659 - val_loss: 0.2733 Epoch 767/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.0728 - val_loss: 0.0663 Epoch 768/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1859 - val_loss: 0.0798 Epoch 769/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1405 - val_loss: 0.1045 Epoch 770/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.5147 - val_loss: 1.3501 Epoch 771/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1034 - val_loss: 0.0671 Epoch 772/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1266 - val_loss: 0.0572 Epoch 773/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1371 - val_loss: 0.1061 Epoch 774/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.3061 - val_loss: 0.0684 Epoch 775/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.0758 - val_loss: 0.0634 Epoch 776/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3489 - val_loss: 0.1306 Epoch 777/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1307 - val_loss: 0.0813 Epoch 778/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1807 - val_loss: 0.1649 Epoch 779/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2237 - val_loss: 1.8152 Epoch 780/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1745 - val_loss: 0.0949 Epoch 781/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2468 - val_loss: 0.1705 Epoch 782/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1137 - val_loss: 0.7684 Epoch 783/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3673 - val_loss: 0.1360 Epoch 784/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.0986 - val_loss: 0.2973 Epoch 785/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2639 - val_loss: 0.3746 Epoch 786/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.1797 - val_loss: 1.7802 Epoch 787/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2062 - val_loss: 0.1072 Epoch 788/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.1067 - val_loss: 0.2442 Epoch 789/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2788 - val_loss: 0.1460 Epoch 790/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1539 - val_loss: 0.6291 Epoch 791/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.1879 - val_loss: 0.0709 Epoch 792/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.4817 - val_loss: 1.3859 Epoch 793/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1143 - val_loss: 0.0844 Epoch 794/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1823 - val_loss: 0.1352 Epoch 795/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2603 - val_loss: 0.1123 Epoch 796/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0954 - val_loss: 0.1319 Epoch 797/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2455 - val_loss: 0.0922 Epoch 798/1000 3888/3888 [==============================] - 0s 120us/sample - loss: 0.1473 - val_loss: 0.0696 Epoch 799/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1855 - val_loss: 0.0917 Epoch 800/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2594 - val_loss: 0.1642 Epoch 801/1000 3888/3888 [==============================] - 0s 123us/sample - loss: 0.1461 - val_loss: 0.0578 Epoch 802/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.4558 - val_loss: 0.0568 Epoch 803/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1232 - val_loss: 0.0516 Epoch 804/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1454 - val_loss: 0.1457 Epoch 805/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.2751 - val_loss: 0.0510 Epoch 806/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1485 - val_loss: 0.1769 Epoch 807/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1418 - val_loss: 0.0852 Epoch 808/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2565 - val_loss: 0.0827 Epoch 809/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1163 - val_loss: 0.1923 Epoch 810/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.2136 - val_loss: 0.1623 Epoch 811/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2176 - val_loss: 0.0700 Epoch 812/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.4820 - val_loss: 0.0732 Epoch 813/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.1121 - val_loss: 0.0541 Epoch 814/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1720 - val_loss: 0.3470 Epoch 815/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1205 - val_loss: 0.1808 Epoch 816/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2112 - val_loss: 0.0697 Epoch 817/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2077 - val_loss: 0.1029 Epoch 818/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3010 - val_loss: 0.2108 Epoch 819/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.0849 - val_loss: 0.0562 Epoch 820/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1624 - val_loss: 0.3852 Epoch 821/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2306 - val_loss: 0.1650 Epoch 822/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.8358 - val_loss: 0.1328 Epoch 823/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.0834 - val_loss: 0.0783 Epoch 824/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2108 - val_loss: 0.0845 Epoch 825/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.0779 - val_loss: 0.0878 Epoch 826/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1430 - val_loss: 0.0663 Epoch 827/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.1883 - val_loss: 0.0892 Epoch 828/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1681 - val_loss: 0.0564 Epoch 829/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3370 - val_loss: 0.1253 Epoch 830/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1476 - val_loss: 0.0696 Epoch 831/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.0802 - val_loss: 0.1129 Epoch 832/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1985 - val_loss: 0.0917 Epoch 833/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3157 - val_loss: 0.5762 Epoch 834/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1232 - val_loss: 0.1640 Epoch 835/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2095 - val_loss: 0.2085 Epoch 836/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2180 - val_loss: 0.1537 Epoch 837/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2545 - val_loss: 0.1015 Epoch 838/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.0856 - val_loss: 0.1532 Epoch 839/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3511 - val_loss: 0.0935 Epoch 840/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2789 - val_loss: 0.4443 Epoch 841/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.0760 - val_loss: 0.0718 Epoch 842/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2409 - val_loss: 0.0730 Epoch 843/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1292 - val_loss: 0.0483 Epoch 844/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2122 - val_loss: 0.0731 Epoch 845/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1212 - val_loss: 0.0755 Epoch 846/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3623 - val_loss: 0.0610 Epoch 847/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.0904 - val_loss: 0.0584 Epoch 848/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.4007 - val_loss: 0.0651 Epoch 849/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1434 - val_loss: 0.2775 Epoch 850/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3900 - val_loss: 0.1619 Epoch 851/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1121 - val_loss: 0.3051 Epoch 852/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1336 - val_loss: 0.0998 Epoch 853/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0942 - val_loss: 0.0700 Epoch 854/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2651 - val_loss: 0.3456 Epoch 855/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1263 - val_loss: 0.0738 Epoch 856/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2104 - val_loss: 0.4709 Epoch 857/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1399 - val_loss: 0.0925 Epoch 858/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.4500 - val_loss: 0.0723 Epoch 859/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.0988 - val_loss: 0.1748 Epoch 860/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1238 - val_loss: 0.0631 Epoch 861/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1344 - val_loss: 0.2982 Epoch 862/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2085 - val_loss: 0.1733 Epoch 863/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2137 - val_loss: 0.2621 Epoch 864/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2446 - val_loss: 0.1470 Epoch 865/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1332 - val_loss: 0.4672 Epoch 866/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.3418 - val_loss: 0.0542 Epoch 867/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1685 - val_loss: 0.0580 Epoch 868/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1020 - val_loss: 0.1579 Epoch 869/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2826 - val_loss: 0.0702 Epoch 870/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1841 - val_loss: 5.3069 Epoch 871/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2151 - val_loss: 0.0502 Epoch 872/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0866 - val_loss: 1.0961 Epoch 873/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3490 - val_loss: 0.1516 Epoch 874/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.1303 - val_loss: 0.1554 Epoch 875/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1617 - val_loss: 0.0894 Epoch 876/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3499 - val_loss: 0.0751 Epoch 877/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.0855 - val_loss: 0.0846 Epoch 878/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.3049 - val_loss: 0.1382 Epoch 879/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0721 - val_loss: 0.0715 Epoch 880/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.5565 - val_loss: 0.2243 Epoch 881/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1206 - val_loss: 0.0680 Epoch 882/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1344 - val_loss: 0.0644 Epoch 883/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2247 - val_loss: 0.0570 Epoch 884/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0919 - val_loss: 0.0868 Epoch 885/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1041 - val_loss: 1.4045 Epoch 886/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1450 - val_loss: 0.1644 Epoch 887/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.4869 - val_loss: 0.0537 Epoch 888/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0761 - val_loss: 0.0876 Epoch 889/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1125 - val_loss: 0.0860 Epoch 890/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2137 - val_loss: 0.0858 Epoch 891/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1387 - val_loss: 0.2571 Epoch 892/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1840 - val_loss: 0.0710 Epoch 893/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2732 - val_loss: 0.0931 Epoch 894/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.0850 - val_loss: 0.0951 Epoch 895/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3204 - val_loss: 0.0789 Epoch 896/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1110 - val_loss: 0.5431 Epoch 897/1000 3888/3888 [==============================] - 0s 112us/sample - loss: 0.2259 - val_loss: 0.1011 Epoch 898/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2424 - val_loss: 0.1982 Epoch 899/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0882 - val_loss: 0.0662 Epoch 900/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.3083 - val_loss: 0.1104 Epoch 901/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2779 - val_loss: 0.0964 Epoch 902/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.0793 - val_loss: 0.0783 Epoch 903/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1236 - val_loss: 0.0902 Epoch 904/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2208 - val_loss: 0.1078 Epoch 905/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1604 - val_loss: 5.0204 Epoch 906/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2848 - val_loss: 0.2597 Epoch 907/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.2453 - val_loss: 0.0750 Epoch 908/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.0978 - val_loss: 0.0637 Epoch 909/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1507 - val_loss: 0.2213 Epoch 910/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.1963 - val_loss: 0.2490 Epoch 911/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1776 - val_loss: 0.1265 Epoch 912/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2867 - val_loss: 0.1448 Epoch 913/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1363 - val_loss: 0.3233 Epoch 914/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2373 - val_loss: 0.1823 Epoch 915/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2536 - val_loss: 0.0495 Epoch 916/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1049 - val_loss: 0.3886 Epoch 917/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1610 - val_loss: 0.2207 Epoch 918/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.2647 - val_loss: 0.6126 Epoch 919/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1382 - val_loss: 0.1665 Epoch 920/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1632 - val_loss: 0.1943 Epoch 921/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1482 - val_loss: 0.5755 Epoch 922/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.2332 - val_loss: 0.0651 Epoch 923/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.2460 - val_loss: 0.2919 Epoch 924/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.0934 - val_loss: 0.0668 Epoch 925/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.6473 - val_loss: 0.0789 Epoch 926/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.0716 - val_loss: 0.1687 Epoch 927/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.0800 - val_loss: 0.8557 Epoch 928/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3238 - val_loss: 0.0722 Epoch 929/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.0649 - val_loss: 0.1097 Epoch 930/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1116 - val_loss: 0.2908 Epoch 931/1000 3888/3888 [==============================] - 0s 117us/sample - loss: 0.1877 - val_loss: 2.2144 Epoch 932/1000 3888/3888 [==============================] - 0s 118us/sample - loss: 0.3020 - val_loss: 0.2213 Epoch 933/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1047 - val_loss: 0.1887 Epoch 934/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.1420 - val_loss: 0.1794 Epoch 935/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.2041 - val_loss: 0.0832 Epoch 936/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.3823 - val_loss: 0.0686 Epoch 937/1000 3888/3888 [==============================] - 0s 115us/sample - loss: 0.0563 - val_loss: 0.0967 Epoch 938/1000 3888/3888 [==============================] - 0s 113us/sample - loss: 0.1263 - val_loss: 0.1219 Epoch 939/1000 3888/3888 [==============================] - 0s 114us/sample - loss: 0.1903 - val_loss: 2.4087 Epoch 940/1000 3888/3888 [==============================] - 0s 119us/sample - loss: 0.3117 - val_loss: 0.0814 Epoch 941/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.3101 - val_loss: 0.1388 Epoch 942/1000 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1111 - val_loss: 0.7598 Epoch 943/1000 3360/3888 [========================>.....] - ETA: 0s - loss: 0.1674Restoring model weights from the end of the best epoch. 3888/3888 [==============================] - 0s 116us/sample - loss: 0.1553 - val_loss: 0.0572 Epoch 00943: early stopping
In [55]:
print(history.history.keys())
print('best value: ', autoencoder.evaluate(X_train_pca, X_train_pca, verbose=0))
pd.DataFrame(history.history).plot(figsize=(8, 5), logy=True)
plt.grid()
dict_keys(['loss', 'val_loss']) best value: 0.04827761372604979
In [56]:
X_reconstructions = autoencoder.predict(X_train_pca)
X_reconstructions = pca.inverse_transform(X_reconstructions)
calculateerror(X_train_1D.reshape(len(times),len(groups),nl,nc),
X_reconstructions.reshape(len(times),len(groups),nl,nc),
groups,
print_step=0)
max_abs_error: 2.195369238521245 mean_abs_error: 0.026052220861696996
/home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: divide by zero encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until /home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until
In [57]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(X_reconstructions.reshape(len(times),len(groups),nl,nc)[100,i,:,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [58]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_train_1D[:,i*nl*nc+4])
ax.flatten()[i].plot(times, X_reconstructions[:,i*nl*nc+4],'--')
ax.flatten()[i].set_title(group)
Convolutional autoencoder¶
In [59]:
#tf.keras.backend.set_image_data_format('channels_first')
tf.keras.backend.image_data_format()
Out[59]:
'channels_last'
In [60]:
#from sklearn.model_selection import train_test_split
#X_train, X_valid = train_test_split(X_train_3D_norm, test_size=0.2, random_state=42)
X_train = np.moveaxis(X_train_3D_norm, 1, 3) # for channel last
In [61]:
tf.random.set_seed(42)
np.random.seed(42)
# Need to have validation loss
early_stopping = keras.callbacks.EarlyStopping(monitor='val_loss',
min_delta=0.0,
patience=100,
verbose=2,
restore_best_weights=True)
conv_encoder = keras.models.Sequential([
#keras.layers.Reshape([28, 28, 1], input_shape=[28, 28]),
keras.layers.InputLayer(input_shape=(10, 10, 8)),
keras.layers.Conv2D(64, kernel_size=3, padding="SAME", activation="elu"),
keras.layers.Flatten(),
keras.layers.Dense(100, activation="elu"),
keras.layers.Dense(50, activation="elu"),
keras.layers.Dense(15)
])
conv_decoder = keras.models.Sequential([
keras.layers.Dense(50, input_shape=[15], activation="elu"),
keras.layers.Dense(100, activation="elu"),
keras.layers.Dense(64*10*10, activation="elu"),
keras.layers.Reshape(target_shape=(10, 10, 64)),
keras.layers.Conv2DTranspose(64, kernel_size=3, strides=1, padding="SAME", activation="elu"),
keras.layers.Conv2DTranspose(8, kernel_size=3, strides=1, padding="SAME"),
keras.layers.Reshape([10, 10, 8])
])
conv_ae = keras.models.Sequential([conv_encoder, conv_decoder])
conv_ae.compile(loss="mse",
optimizer=keras.optimizers.Nadam(lr=0.0001, beta_1=0.9, beta_2=0.999))
conv_encoder.summary()
conv_decoder.summary()
Model: "sequential_12" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= conv2d (Conv2D) (None, 10, 10, 64) 4672 _________________________________________________________________ flatten (Flatten) (None, 6400) 0 _________________________________________________________________ dense_18 (Dense) (None, 100) 640100 _________________________________________________________________ dense_19 (Dense) (None, 50) 5050 _________________________________________________________________ dense_20 (Dense) (None, 15) 765 ================================================================= Total params: 650,587 Trainable params: 650,587 Non-trainable params: 0 _________________________________________________________________ Model: "sequential_13" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= dense_21 (Dense) (None, 50) 800 _________________________________________________________________ dense_22 (Dense) (None, 100) 5100 _________________________________________________________________ dense_23 (Dense) (None, 6400) 646400 _________________________________________________________________ reshape (Reshape) (None, 10, 10, 64) 0 _________________________________________________________________ conv2d_transpose (Conv2DTran (None, 10, 10, 64) 36928 _________________________________________________________________ conv2d_transpose_1 (Conv2DTr (None, 10, 10, 8) 4616 _________________________________________________________________ reshape_1 (Reshape) (None, 10, 10, 8) 0 ================================================================= Total params: 693,844 Trainable params: 693,844 Non-trainable params: 0 _________________________________________________________________
In [62]:
history = conv_ae.fit(X_train, X_train,
epochs=1000,
validation_data=(X_train, X_train),
callbacks=[early_stopping])
Train on 3888 samples, validate on 3888 samples Epoch 1/1000 3888/3888 [==============================] - 3s 895us/sample - loss: 0.0434 - val_loss: 0.0080 Epoch 2/1000 3888/3888 [==============================] - 2s 558us/sample - loss: 0.0036 - val_loss: 0.0015 Epoch 3/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 0.0011 - val_loss: 7.2160e-04 Epoch 4/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 6.3083e-04 - val_loss: 4.0945e-04 Epoch 5/1000 3888/3888 [==============================] - 2s 538us/sample - loss: 6.5230e-04 - val_loss: 3.1501e-04 Epoch 6/1000 3888/3888 [==============================] - 2s 535us/sample - loss: 3.0039e-04 - val_loss: 2.3431e-04 Epoch 7/1000 3888/3888 [==============================] - 2s 539us/sample - loss: 3.5650e-04 - val_loss: 1.9783e-04 Epoch 8/1000 3888/3888 [==============================] - 2s 533us/sample - loss: 2.3832e-04 - val_loss: 1.7775e-04 Epoch 9/1000 3888/3888 [==============================] - 2s 556us/sample - loss: 2.5771e-04 - val_loss: 1.7263e-04 Epoch 10/1000 3888/3888 [==============================] - 2s 537us/sample - loss: 2.9760e-04 - val_loss: 3.2514e-04 Epoch 11/1000 3888/3888 [==============================] - 2s 544us/sample - loss: 1.6701e-04 - val_loss: 1.2307e-04 Epoch 12/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 1.8069e-04 - val_loss: 2.7333e-04 Epoch 13/1000 3888/3888 [==============================] - 2s 543us/sample - loss: 2.5737e-04 - val_loss: 8.7746e-05 Epoch 14/1000 3888/3888 [==============================] - 2s 640us/sample - loss: 1.2636e-04 - val_loss: 6.6394e-04 Epoch 15/1000 3888/3888 [==============================] - 2s 581us/sample - loss: 1.7777e-04 - val_loss: 8.4140e-05 Epoch 16/1000 3888/3888 [==============================] - 2s 549us/sample - loss: 2.2911e-04 - val_loss: 7.2874e-05 Epoch 17/1000 3888/3888 [==============================] - 2s 546us/sample - loss: 1.3529e-04 - val_loss: 7.1527e-05 Epoch 18/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 9.6322e-05 - val_loss: 9.0245e-05 Epoch 19/1000 3888/3888 [==============================] - 2s 542us/sample - loss: 1.5794e-04 - val_loss: 6.4697e-05 Epoch 20/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 1.0102e-04 - val_loss: 7.2958e-05 Epoch 21/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 1.6049e-04 - val_loss: 5.9588e-05 Epoch 22/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 9.6004e-05 - val_loss: 4.6323e-05 Epoch 23/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 1.2193e-04 - val_loss: 7.0513e-05 Epoch 24/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 9.5553e-05 - val_loss: 4.6434e-05 Epoch 25/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 1.3025e-04 - val_loss: 6.7973e-05 Epoch 26/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 7.9251e-05 - val_loss: 3.7723e-04 Epoch 27/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 7.1520e-05 - val_loss: 4.5085e-05 Epoch 28/1000 3888/3888 [==============================] - 2s 538us/sample - loss: 1.5674e-04 - val_loss: 8.4231e-05 Epoch 29/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 6.7590e-05 - val_loss: 5.9001e-05 Epoch 30/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 9.6203e-05 - val_loss: 4.4551e-04 Epoch 31/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 6.8562e-05 - val_loss: 5.6724e-05 Epoch 32/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.1168e-04 - val_loss: 2.8260e-05 Epoch 33/1000 3888/3888 [==============================] - 2s 596us/sample - loss: 1.9677e-04 - val_loss: 3.2629e-05 Epoch 34/1000 3888/3888 [==============================] - 2s 562us/sample - loss: 3.0406e-05 - val_loss: 4.4957e-05 Epoch 35/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 4.0286e-05 - val_loss: 3.0143e-05 Epoch 36/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 9.4810e-05 - val_loss: 3.9579e-05 Epoch 37/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 3.7689e-05 - val_loss: 3.2112e-05 Epoch 38/1000 3888/3888 [==============================] - 2s 503us/sample - loss: 1.5882e-04 - val_loss: 4.4055e-05 Epoch 39/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 3.0080e-05 - val_loss: 2.1989e-05 Epoch 40/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 3.5845e-05 - val_loss: 3.1032e-05 Epoch 41/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 4.5945e-05 - val_loss: 2.4032e-05 Epoch 42/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 6.9854e-05 - val_loss: 0.0019 Epoch 43/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 1.2242e-04 - val_loss: 5.3096e-05 Epoch 44/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 9.9203e-05 - val_loss: 2.8848e-05 Epoch 45/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 2.2306e-05 - val_loss: 1.8452e-05 Epoch 46/1000 3888/3888 [==============================] - 2s 508us/sample - loss: 4.7209e-05 - val_loss: 2.0237e-05 Epoch 47/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 3.3210e-05 - val_loss: 2.8410e-05 Epoch 48/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 4.7201e-05 - val_loss: 2.2302e-05 Epoch 49/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.7401e-04 - val_loss: 1.7304e-05 Epoch 50/1000 3888/3888 [==============================] - 2s 535us/sample - loss: 1.9281e-05 - val_loss: 2.1217e-05 Epoch 51/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 4.9003e-05 - val_loss: 2.4669e-05 Epoch 52/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 2.1840e-05 - val_loss: 2.0712e-05 Epoch 53/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 3.2953e-05 - val_loss: 8.5942e-05 Epoch 54/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 5.3268e-05 - val_loss: 1.5969e-05 Epoch 55/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 1.4217e-04 - val_loss: 1.5205e-05 Epoch 56/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 1.7671e-05 - val_loss: 1.8281e-05 Epoch 57/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 2.3153e-05 - val_loss: 2.6244e-05 Epoch 58/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 2.6138e-05 - val_loss: 6.9503e-04 Epoch 59/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 4.4097e-05 - val_loss: 1.9275e-05 Epoch 60/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 1.9835e-05 - val_loss: 1.2806e-05 Epoch 61/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 3.8098e-05 - val_loss: 3.2059e-05 Epoch 62/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 7.4090e-05 - val_loss: 3.2743e-05 Epoch 63/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 2.0525e-05 - val_loss: 1.6393e-05 Epoch 64/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 2.5792e-05 - val_loss: 1.4576e-05 Epoch 65/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 3.6762e-05 - val_loss: 1.7539e-05 Epoch 66/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 3.7106e-05 - val_loss: 1.4789e-05 Epoch 67/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.3481e-04 - val_loss: 9.9734e-05 Epoch 68/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 1.8228e-05 - val_loss: 1.2812e-05 Epoch 69/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 2.3419e-05 - val_loss: 1.8164e-05 Epoch 70/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 2.2865e-05 - val_loss: 1.2283e-05 Epoch 71/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.1961e-05 - val_loss: 1.1199e-05 Epoch 72/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.5066e-05 - val_loss: 1.4087e-05 Epoch 73/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 3.0623e-05 - val_loss: 2.6909e-05 Epoch 74/1000 3888/3888 [==============================] - 2s 510us/sample - loss: 2.5317e-05 - val_loss: 1.0060e-05 Epoch 75/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 6.0347e-05 - val_loss: 1.3934e-05 Epoch 76/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.4041e-05 - val_loss: 1.2806e-05 Epoch 77/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 2.6030e-05 - val_loss: 1.6550e-05 Epoch 78/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 3.3747e-05 - val_loss: 9.9822e-06 Epoch 79/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.7553e-05 - val_loss: 1.7042e-04 Epoch 80/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 4.5654e-05 - val_loss: 2.2506e-05 Epoch 81/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 1.9163e-05 - val_loss: 1.1224e-05 Epoch 82/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.6265e-05 - val_loss: 1.7619e-05 Epoch 83/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 4.3717e-05 - val_loss: 1.4063e-05 Epoch 84/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 3.5289e-05 - val_loss: 9.6553e-06 Epoch 85/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 1.4116e-05 - val_loss: 2.7602e-05 Epoch 86/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 2.5642e-05 - val_loss: 6.0762e-05 Epoch 87/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 3.9845e-05 - val_loss: 1.5329e-05 Epoch 88/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 2.6384e-05 - val_loss: 2.9309e-04 Epoch 89/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 2.0632e-05 - val_loss: 2.0136e-05 Epoch 90/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 1.4065e-05 - val_loss: 1.8223e-05 Epoch 91/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 3.5978e-05 - val_loss: 1.1969e-05 Epoch 92/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 2.0140e-05 - val_loss: 1.3532e-05 Epoch 93/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.4949e-05 - val_loss: 7.4074e-06 Epoch 94/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 1.6803e-05 - val_loss: 1.9807e-05 Epoch 95/1000 3888/3888 [==============================] - 2s 541us/sample - loss: 4.0814e-05 - val_loss: 2.1695e-05 Epoch 96/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.1174e-05 - val_loss: 1.5759e-05 Epoch 97/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 1.7318e-05 - val_loss: 3.7607e-05 Epoch 98/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.9572e-05 - val_loss: 1.1790e-05 Epoch 99/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 3.9336e-05 - val_loss: 3.8419e-05 Epoch 100/1000 3888/3888 [==============================] - 2s 510us/sample - loss: 1.4113e-05 - val_loss: 1.1696e-05 Epoch 101/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 4.7618e-05 - val_loss: 2.4654e-05 Epoch 102/1000 3888/3888 [==============================] - 2s 534us/sample - loss: 1.0097e-05 - val_loss: 8.5651e-06 Epoch 103/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 8.5207e-06 - val_loss: 4.1195e-05 Epoch 104/1000 3888/3888 [==============================] - 2s 510us/sample - loss: 1.9288e-05 - val_loss: 3.5934e-05 Epoch 105/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.4522e-05 - val_loss: 4.7873e-05 Epoch 106/1000 3888/3888 [==============================] - 2s 510us/sample - loss: 2.8465e-05 - val_loss: 8.8930e-06 Epoch 107/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.2106e-05 - val_loss: 8.2797e-06 Epoch 108/1000 3888/3888 [==============================] - 2s 506us/sample - loss: 2.1212e-05 - val_loss: 1.3833e-05 Epoch 109/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 3.1905e-05 - val_loss: 7.8618e-06 Epoch 110/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 1.2219e-05 - val_loss: 1.3177e-05 Epoch 111/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 3.6925e-05 - val_loss: 8.0115e-06 Epoch 112/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 6.7165e-06 - val_loss: 8.4891e-06 Epoch 113/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 1.1984e-05 - val_loss: 2.1423e-05 Epoch 114/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.8656e-05 - val_loss: 1.8823e-05 Epoch 115/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.1904e-05 - val_loss: 8.2761e-06 Epoch 116/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 4.0860e-05 - val_loss: 5.4605e-05 Epoch 117/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 8.4449e-06 - val_loss: 5.1097e-06 Epoch 118/1000 3888/3888 [==============================] - 2s 506us/sample - loss: 7.8520e-06 - val_loss: 2.2092e-05 Epoch 119/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.5702e-05 - val_loss: 4.5121e-05 Epoch 120/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 8.6415e-06 - val_loss: 7.5840e-06 Epoch 121/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.3241e-05 - val_loss: 5.7292e-04 Epoch 122/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 2.6270e-05 - val_loss: 6.8134e-06 Epoch 123/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 8.2742e-06 - val_loss: 1.1918e-05 Epoch 124/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.3204e-05 - val_loss: 9.0442e-06 Epoch 125/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 1.7324e-05 - val_loss: 7.0954e-06 Epoch 126/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.4779e-05 - val_loss: 5.9253e-06 Epoch 127/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 2.4810e-05 - val_loss: 9.5395e-05 Epoch 128/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 1.5551e-05 - val_loss: 7.6527e-06 Epoch 129/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 2.2809e-05 - val_loss: 7.3738e-06 Epoch 130/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 5.9714e-06 - val_loss: 8.6370e-06 Epoch 131/1000 3888/3888 [==============================] - 2s 536us/sample - loss: 2.4162e-05 - val_loss: 1.0039e-05 Epoch 132/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 9.6565e-06 - val_loss: 6.6872e-06 Epoch 133/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 1.0361e-05 - val_loss: 8.5798e-05 Epoch 134/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.2593e-05 - val_loss: 5.7562e-06 Epoch 135/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 5.7494e-06 - val_loss: 1.4141e-05 Epoch 136/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 3.5733e-05 - val_loss: 3.2080e-05 Epoch 137/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 6.8600e-06 - val_loss: 7.5116e-06 Epoch 138/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.2544e-05 - val_loss: 1.3008e-05 Epoch 139/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.5438e-05 - val_loss: 7.3079e-06 Epoch 140/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 1.4497e-05 - val_loss: 1.0710e-05 Epoch 141/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 8.9607e-06 - val_loss: 2.4513e-05 Epoch 142/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 8.4252e-06 - val_loss: 5.6528e-06 Epoch 143/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.5146e-05 - val_loss: 5.6560e-06 Epoch 144/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.2827e-05 - val_loss: 2.4622e-05 Epoch 145/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 2.6252e-05 - val_loss: 7.7830e-06 Epoch 146/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 9.8901e-06 - val_loss: 7.7425e-06 Epoch 147/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 1.0156e-05 - val_loss: 6.5091e-06 Epoch 148/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 2.2889e-05 - val_loss: 7.4927e-06 Epoch 149/1000 3888/3888 [==============================] - 2s 535us/sample - loss: 1.2111e-05 - val_loss: 5.6396e-06 Epoch 150/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 1.1303e-05 - val_loss: 1.0146e-05 Epoch 151/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 1.1524e-05 - val_loss: 1.0757e-04 Epoch 152/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 1.2155e-05 - val_loss: 7.1149e-06 Epoch 153/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 1.9844e-05 - val_loss: 4.1217e-06 Epoch 154/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 7.2853e-06 - val_loss: 2.3351e-05 Epoch 155/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 1.3743e-05 - val_loss: 5.8358e-06 Epoch 156/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.1266e-05 - val_loss: 9.5893e-05 Epoch 157/1000 3888/3888 [==============================] - 2s 537us/sample - loss: 3.9770e-05 - val_loss: 4.6877e-06 Epoch 158/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 4.2406e-06 - val_loss: 4.8267e-06 Epoch 159/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 6.5291e-06 - val_loss: 7.1137e-06 Epoch 160/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 5.1563e-06 - val_loss: 1.2186e-05 Epoch 161/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 1.1969e-05 - val_loss: 1.5704e-05 Epoch 162/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 1.1371e-05 - val_loss: 7.0800e-06 Epoch 163/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.1869e-05 - val_loss: 5.0365e-06 Epoch 164/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 1.7494e-05 - val_loss: 5.4833e-06 Epoch 165/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 8.1931e-06 - val_loss: 1.6637e-05 Epoch 166/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 4.8346e-05 - val_loss: 4.1029e-06 Epoch 167/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 3.8181e-06 - val_loss: 4.1837e-06 Epoch 168/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 4.1776e-06 - val_loss: 4.5184e-06 Epoch 169/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 6.1066e-06 - val_loss: 4.4452e-06 Epoch 170/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 7.8181e-06 - val_loss: 5.5000e-06 Epoch 171/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.0267e-05 - val_loss: 6.7189e-06 Epoch 172/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 8.2655e-06 - val_loss: 6.9044e-05 Epoch 173/1000 3888/3888 [==============================] - 2s 505us/sample - loss: 1.1067e-05 - val_loss: 1.3984e-05 Epoch 174/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 2.6084e-05 - val_loss: 6.4030e-06 Epoch 175/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 5.6664e-06 - val_loss: 1.0517e-05 Epoch 176/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 4.3920e-06 - val_loss: 1.0428e-05 Epoch 177/1000 3888/3888 [==============================] - 2s 509us/sample - loss: 2.3571e-05 - val_loss: 5.7982e-06 Epoch 178/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 4.8672e-06 - val_loss: 8.1345e-06 Epoch 179/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 1.0596e-05 - val_loss: 1.8635e-05 Epoch 180/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 1.0320e-05 - val_loss: 3.8052e-06 Epoch 181/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 9.6965e-06 - val_loss: 7.7871e-06 Epoch 182/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.3212e-05 - val_loss: 8.9347e-06 Epoch 183/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 5.3752e-06 - val_loss: 1.2881e-05 Epoch 184/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.0595e-05 - val_loss: 3.3365e-06 Epoch 185/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 1.3169e-05 - val_loss: 7.2238e-06 Epoch 186/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 9.8885e-06 - val_loss: 9.1990e-06 Epoch 187/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 1.5256e-05 - val_loss: 5.4908e-06 Epoch 188/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.6004e-05 - val_loss: 4.8164e-06 Epoch 189/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 4.9121e-06 - val_loss: 6.4997e-06 Epoch 190/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 8.0949e-06 - val_loss: 4.1455e-05 Epoch 191/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 9.4887e-06 - val_loss: 6.6267e-06 Epoch 192/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 1.1911e-05 - val_loss: 7.8412e-06 Epoch 193/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 2.2724e-05 - val_loss: 3.9957e-06 Epoch 194/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 4.1641e-06 - val_loss: 3.6641e-06 Epoch 195/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.6077e-05 - val_loss: 4.9613e-06 Epoch 196/1000 3888/3888 [==============================] - 2s 533us/sample - loss: 4.1301e-06 - val_loss: 3.1753e-06 Epoch 197/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 1.3887e-05 - val_loss: 4.2216e-06 Epoch 198/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 4.2333e-06 - val_loss: 9.9052e-05 Epoch 199/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.7420e-05 - val_loss: 6.1529e-06 Epoch 200/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 6.3075e-06 - val_loss: 6.4819e-06 Epoch 201/1000 3888/3888 [==============================] - 2s 507us/sample - loss: 6.9023e-06 - val_loss: 1.9570e-05 Epoch 202/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.2040e-05 - val_loss: 5.9059e-06 Epoch 203/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 3.1139e-05 - val_loss: 8.3588e-06 Epoch 204/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 3.9011e-06 - val_loss: 3.5444e-06 Epoch 205/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 3.1178e-06 - val_loss: 3.4576e-06 Epoch 206/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 9.0759e-06 - val_loss: 3.6263e-06 Epoch 207/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 7.6604e-06 - val_loss: 9.3612e-06 Epoch 208/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 7.4209e-06 - val_loss: 7.7182e-06 Epoch 209/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 7.0581e-06 - val_loss: 1.0622e-04 Epoch 210/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 1.1246e-05 - val_loss: 7.2725e-06 Epoch 211/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.1250e-05 - val_loss: 8.1946e-05 Epoch 212/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 1.2572e-05 - val_loss: 4.7253e-06 Epoch 213/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 7.2925e-06 - val_loss: 4.5228e-06 Epoch 214/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.3443e-05 - val_loss: 3.5181e-06 Epoch 215/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 4.9517e-06 - val_loss: 1.3186e-05 Epoch 216/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.9197e-05 - val_loss: 4.1599e-06 Epoch 217/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 4.9010e-06 - val_loss: 5.7298e-06 Epoch 218/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 6.2516e-06 - val_loss: 5.1742e-06 Epoch 219/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 6.0227e-05 - val_loss: 6.2305e-05 Epoch 220/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 8.5514e-06 - val_loss: 8.4053e-06 Epoch 221/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 2.9974e-06 - val_loss: 3.1485e-06 Epoch 222/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 8.2789e-06 - val_loss: 2.7642e-06 Epoch 223/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 5.3145e-06 - val_loss: 1.4580e-05 Epoch 224/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 4.6759e-06 - val_loss: 4.4093e-06 Epoch 225/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 2.8738e-05 - val_loss: 2.3206e-05 Epoch 226/1000 3888/3888 [==============================] - 2s 534us/sample - loss: 3.8442e-06 - val_loss: 3.3573e-06 Epoch 227/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 3.5618e-06 - val_loss: 2.7529e-06 Epoch 228/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 3.3082e-06 - val_loss: 4.1051e-06 Epoch 229/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 7.8566e-06 - val_loss: 2.0916e-05 Epoch 230/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.5488e-05 - val_loss: 3.9161e-06 Epoch 231/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 6.2693e-06 - val_loss: 1.0627e-05 Epoch 232/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 5.1284e-06 - val_loss: 1.1321e-05 Epoch 233/1000 3888/3888 [==============================] - 2s 536us/sample - loss: 8.6836e-06 - val_loss: 3.2489e-05 Epoch 234/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.1587e-05 - val_loss: 2.7507e-06 Epoch 235/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 6.7578e-06 - val_loss: 5.5140e-06 Epoch 236/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 7.0778e-06 - val_loss: 2.3809e-05 Epoch 237/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 1.3034e-05 - val_loss: 3.9254e-06 Epoch 238/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 7.1849e-06 - val_loss: 4.3003e-06 Epoch 239/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 8.5896e-06 - val_loss: 1.2718e-05 Epoch 240/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 1.3826e-05 - val_loss: 2.8758e-06 Epoch 241/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 4.1601e-06 - val_loss: 3.8652e-06 Epoch 242/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 6.1050e-06 - val_loss: 1.1409e-05 Epoch 243/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 8.8223e-06 - val_loss: 4.1144e-06 Epoch 244/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 8.8962e-06 - val_loss: 3.4990e-06 Epoch 245/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 1.4779e-05 - val_loss: 4.1267e-06 Epoch 246/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 5.9276e-06 - val_loss: 7.9941e-06 Epoch 247/1000 3888/3888 [==============================] - 2s 507us/sample - loss: 7.0969e-06 - val_loss: 2.7558e-06 Epoch 248/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 2.2623e-05 - val_loss: 5.4240e-06 Epoch 249/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 3.0067e-06 - val_loss: 2.3855e-05 Epoch 250/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.3376e-05 - val_loss: 4.7684e-06 Epoch 251/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 3.1349e-06 - val_loss: 2.7506e-06 Epoch 252/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 3.0233e-06 - val_loss: 2.7614e-06 Epoch 253/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 5.4205e-06 - val_loss: 5.4551e-06 Epoch 254/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 9.0579e-06 - val_loss: 1.3623e-04 Epoch 255/1000 3888/3888 [==============================] - 2s 509us/sample - loss: 1.1037e-05 - val_loss: 3.1331e-06 Epoch 256/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 5.8418e-06 - val_loss: 1.3900e-05 Epoch 257/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 2.0149e-05 - val_loss: 5.6909e-06 Epoch 258/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 3.1772e-06 - val_loss: 3.9846e-06 Epoch 259/1000 3888/3888 [==============================] - 2s 509us/sample - loss: 3.8305e-06 - val_loss: 6.8447e-06 Epoch 260/1000 3888/3888 [==============================] - 2s 504us/sample - loss: 9.0661e-06 - val_loss: 3.3400e-06 Epoch 261/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 6.0070e-06 - val_loss: 5.0809e-06 Epoch 262/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 5.7080e-06 - val_loss: 4.8959e-06 Epoch 263/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 9.4947e-06 - val_loss: 5.6802e-06 Epoch 264/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 1.1174e-05 - val_loss: 5.8145e-06 Epoch 265/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 9.3153e-06 - val_loss: 5.9804e-06 Epoch 266/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 6.6430e-06 - val_loss: 2.1847e-06 Epoch 267/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 8.5639e-06 - val_loss: 6.0531e-06 Epoch 268/1000 3888/3888 [==============================] - 2s 507us/sample - loss: 1.6147e-05 - val_loss: 3.4445e-05 Epoch 269/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 4.4172e-06 - val_loss: 3.6505e-06 Epoch 270/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 3.5948e-06 - val_loss: 5.2941e-06 Epoch 271/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 1.9204e-05 - val_loss: 4.1708e-06 Epoch 272/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 3.0636e-06 - val_loss: 2.8033e-06 Epoch 273/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 3.7931e-06 - val_loss: 1.1575e-05 Epoch 274/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.0040e-05 - val_loss: 5.1195e-06 Epoch 275/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 1.0341e-05 - val_loss: 3.9013e-06 Epoch 276/1000 3888/3888 [==============================] - 2s 508us/sample - loss: 1.3690e-05 - val_loss: 2.3333e-05 Epoch 277/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 4.2282e-06 - val_loss: 2.8866e-06 Epoch 278/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 3.8540e-06 - val_loss: 2.2839e-06 Epoch 279/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 6.1632e-06 - val_loss: 6.5055e-06 Epoch 280/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.2697e-05 - val_loss: 9.6706e-06 Epoch 281/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 5.0937e-06 - val_loss: 4.4150e-06 Epoch 282/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 9.1255e-06 - val_loss: 2.6147e-06 Epoch 283/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.1346e-05 - val_loss: 1.4970e-05 Epoch 284/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.9381e-05 - val_loss: 2.8766e-06 Epoch 285/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 2.3445e-06 - val_loss: 2.3841e-06 Epoch 286/1000 3888/3888 [==============================] - 2s 509us/sample - loss: 2.9023e-06 - val_loss: 5.4774e-06 Epoch 287/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 5.9073e-06 - val_loss: 2.5260e-06 Epoch 288/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 5.0261e-06 - val_loss: 4.8703e-06 Epoch 289/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 8.6048e-06 - val_loss: 2.0882e-05 Epoch 290/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 8.0509e-06 - val_loss: 5.8226e-06 Epoch 291/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 5.8337e-06 - val_loss: 3.3843e-06 Epoch 292/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 8.6691e-06 - val_loss: 1.2504e-05 Epoch 293/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 9.5804e-06 - val_loss: 1.2868e-05 Epoch 294/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 8.6848e-06 - val_loss: 6.9553e-06 Epoch 295/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 9.7696e-06 - val_loss: 9.8989e-06 Epoch 296/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 1.1448e-05 - val_loss: 1.0150e-05 Epoch 297/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 4.8309e-06 - val_loss: 3.2402e-06 Epoch 298/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 6.0955e-06 - val_loss: 4.1831e-06 Epoch 299/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 2.0048e-05 - val_loss: 6.8533e-06 Epoch 300/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 2.5143e-06 - val_loss: 1.8474e-06 Epoch 301/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 7.3177e-06 - val_loss: 1.1143e-05 Epoch 302/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 4.9378e-06 - val_loss: 5.8322e-06 Epoch 303/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 4.2476e-06 - val_loss: 2.5494e-05 Epoch 304/1000 3888/3888 [==============================] - 2s 540us/sample - loss: 1.4943e-05 - val_loss: 4.1072e-06 Epoch 305/1000 3888/3888 [==============================] - 2s 534us/sample - loss: 3.2359e-06 - val_loss: 2.9837e-06 Epoch 306/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 4.7132e-06 - val_loss: 7.2627e-06 Epoch 307/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 5.1347e-06 - val_loss: 2.4915e-06 Epoch 308/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.0642e-05 - val_loss: 5.9635e-06 Epoch 309/1000 3888/3888 [==============================] - 2s 535us/sample - loss: 2.2651e-05 - val_loss: 4.8468e-06 Epoch 310/1000 3888/3888 [==============================] - 2s 533us/sample - loss: 2.1439e-06 - val_loss: 2.5291e-06 Epoch 311/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 1.0768e-05 - val_loss: 4.4883e-06 Epoch 312/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 2.3849e-06 - val_loss: 3.5999e-06 Epoch 313/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 5.8903e-06 - val_loss: 3.3549e-06 Epoch 314/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 4.0148e-06 - val_loss: 4.0536e-06 Epoch 315/1000 3888/3888 [==============================] - 2s 533us/sample - loss: 7.4665e-06 - val_loss: 4.3147e-05 Epoch 316/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 8.9976e-06 - val_loss: 2.5554e-06 Epoch 317/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 6.1528e-06 - val_loss: 2.2465e-06 Epoch 318/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 5.4038e-06 - val_loss: 4.1070e-06 Epoch 319/1000 3888/3888 [==============================] - 2s 543us/sample - loss: 4.5286e-06 - val_loss: 5.4174e-06 Epoch 320/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 1.3519e-05 - val_loss: 2.1794e-06 Epoch 321/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 3.5707e-06 - val_loss: 3.5886e-06 Epoch 322/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 8.5194e-06 - val_loss: 8.8490e-06 Epoch 323/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 7.8801e-06 - val_loss: 6.4011e-05 Epoch 324/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 9.5432e-06 - val_loss: 5.1333e-06 Epoch 325/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 4.6362e-06 - val_loss: 3.8785e-06 Epoch 326/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 7.1403e-06 - val_loss: 2.4805e-06 Epoch 327/1000 3888/3888 [==============================] - 2s 510us/sample - loss: 6.4191e-06 - val_loss: 3.9879e-06 Epoch 328/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 6.3707e-06 - val_loss: 2.0556e-05 Epoch 329/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 7.7769e-06 - val_loss: 2.5108e-05 Epoch 330/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 8.6417e-06 - val_loss: 2.8069e-06 Epoch 331/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 5.6465e-06 - val_loss: 1.1017e-05 Epoch 332/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 6.1531e-06 - val_loss: 3.1217e-06 Epoch 333/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 5.7912e-06 - val_loss: 7.2552e-06 Epoch 334/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 8.9585e-06 - val_loss: 1.3419e-05 Epoch 335/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 8.7274e-06 - val_loss: 0.0014 Epoch 336/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 3.7067e-05 - val_loss: 1.8333e-06 Epoch 337/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.8976e-06 - val_loss: 2.1979e-06 Epoch 338/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 2.0615e-06 - val_loss: 2.0269e-06 Epoch 339/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 4.2314e-06 - val_loss: 8.4085e-06 Epoch 340/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 3.2186e-06 - val_loss: 2.9493e-06 Epoch 341/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.3227e-05 - val_loss: 4.1129e-06 Epoch 342/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 2.8331e-06 - val_loss: 4.3980e-06 Epoch 343/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 2.5330e-06 - val_loss: 4.5573e-06 Epoch 344/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 8.7895e-06 - val_loss: 2.0715e-06 Epoch 345/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 3.4397e-06 - val_loss: 3.1976e-06 Epoch 346/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 8.6552e-06 - val_loss: 6.4671e-06 Epoch 347/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 4.9543e-06 - val_loss: 2.4473e-06 Epoch 348/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 8.2030e-06 - val_loss: 2.1625e-05 Epoch 349/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 5.2965e-06 - val_loss: 1.3248e-05 Epoch 350/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 3.5001e-06 - val_loss: 3.4564e-06 Epoch 351/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.1563e-05 - val_loss: 2.1468e-06 Epoch 352/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 4.2157e-06 - val_loss: 2.5329e-06 Epoch 353/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 3.2324e-06 - val_loss: 2.3022e-06 Epoch 354/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 3.0538e-06 - val_loss: 2.3155e-06 Epoch 355/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 9.7924e-06 - val_loss: 1.3314e-05 Epoch 356/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 5.1538e-06 - val_loss: 6.7261e-06 Epoch 357/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 3.9929e-06 - val_loss: 3.2329e-06 Epoch 358/1000 3888/3888 [==============================] - 2s 541us/sample - loss: 6.0119e-06 - val_loss: 1.1234e-05 Epoch 359/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 1.4028e-05 - val_loss: 5.4172e-06 Epoch 360/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 4.1374e-06 - val_loss: 2.3146e-06 Epoch 361/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 8.2467e-06 - val_loss: 1.4046e-05 Epoch 362/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 7.3327e-06 - val_loss: 3.9584e-05 Epoch 363/1000 3888/3888 [==============================] - 2s 538us/sample - loss: 5.9798e-06 - val_loss: 2.4901e-06 Epoch 364/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 1.1299e-05 - val_loss: 3.2310e-06 Epoch 365/1000 3888/3888 [==============================] - 2s 507us/sample - loss: 2.5684e-06 - val_loss: 1.9659e-06 Epoch 366/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 3.7174e-06 - val_loss: 2.2971e-06 Epoch 367/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.3376e-05 - val_loss: 2.3031e-05 Epoch 368/1000 3888/3888 [==============================] - 2s 534us/sample - loss: 2.8427e-06 - val_loss: 2.1986e-06 Epoch 369/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 1.0153e-05 - val_loss: 1.9796e-05 Epoch 370/1000 3888/3888 [==============================] - 2s 534us/sample - loss: 6.7771e-06 - val_loss: 3.6131e-06 Epoch 371/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 1.9891e-06 - val_loss: 2.2860e-06 Epoch 372/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.7958e-06 - val_loss: 2.4769e-06 Epoch 373/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 1.1905e-05 - val_loss: 1.9560e-06 Epoch 374/1000 3888/3888 [==============================] - 2s 533us/sample - loss: 2.6675e-06 - val_loss: 2.0582e-06 Epoch 375/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 2.7737e-06 - val_loss: 5.2559e-06 Epoch 376/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 6.8492e-06 - val_loss: 2.9664e-06 Epoch 377/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 4.7928e-06 - val_loss: 2.6107e-06 Epoch 378/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 9.8566e-06 - val_loss: 1.3531e-05 Epoch 379/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 5.3507e-06 - val_loss: 6.0627e-06 Epoch 380/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 5.5328e-06 - val_loss: 2.2183e-05 Epoch 381/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.0152e-05 - val_loss: 3.3916e-06 Epoch 382/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 8.1791e-06 - val_loss: 2.2890e-06 Epoch 383/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 1.3531e-05 - val_loss: 1.1812e-05 Epoch 384/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 3.2317e-06 - val_loss: 2.0314e-06 Epoch 385/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 3.5770e-06 - val_loss: 3.1171e-05 Epoch 386/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 5.1324e-06 - val_loss: 6.7801e-06 Epoch 387/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 9.9639e-06 - val_loss: 1.2016e-05 Epoch 388/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 8.0675e-06 - val_loss: 1.9771e-06 Epoch 389/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 2.4551e-06 - val_loss: 2.4244e-06 Epoch 390/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 5.0200e-06 - val_loss: 3.2957e-06 Epoch 391/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 4.1900e-06 - val_loss: 1.9228e-06 Epoch 392/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 8.0845e-06 - val_loss: 1.9572e-05 Epoch 393/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 1.3645e-05 - val_loss: 2.6554e-06 Epoch 394/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 4.8637e-06 - val_loss: 2.7313e-06 Epoch 395/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 2.7640e-06 - val_loss: 1.0936e-05 Epoch 396/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 4.7181e-05 - val_loss: 6.4502e-05 Epoch 397/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 5.2762e-06 - val_loss: 2.2628e-06 Epoch 398/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.4881e-06 - val_loss: 1.5366e-06 Epoch 399/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.5426e-06 - val_loss: 1.5187e-06 Epoch 400/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.5232e-06 - val_loss: 1.5205e-06 Epoch 401/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 2.0821e-06 - val_loss: 2.5434e-06 Epoch 402/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 3.4016e-06 - val_loss: 2.6107e-06 Epoch 403/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 5.3434e-06 - val_loss: 7.2095e-06 Epoch 404/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 9.9547e-06 - val_loss: 2.9351e-06 Epoch 405/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 6.2887e-06 - val_loss: 2.3486e-06 Epoch 406/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 5.7482e-06 - val_loss: 2.5819e-06 Epoch 407/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 2.9476e-06 - val_loss: 4.0715e-05 Epoch 408/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 7.3017e-06 - val_loss: 1.4303e-06 Epoch 409/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 2.6444e-06 - val_loss: 2.9202e-05 Epoch 410/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 9.7810e-06 - val_loss: 2.4417e-06 Epoch 411/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 2.8718e-06 - val_loss: 7.4957e-06 Epoch 412/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 6.8378e-06 - val_loss: 5.9957e-06 Epoch 413/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 5.3566e-06 - val_loss: 1.9532e-06 Epoch 414/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 3.7647e-06 - val_loss: 2.8118e-05 Epoch 415/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 6.1785e-06 - val_loss: 6.7588e-05 Epoch 416/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 8.6879e-06 - val_loss: 1.5326e-05 Epoch 417/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 1.3432e-05 - val_loss: 3.4263e-06 Epoch 418/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.4245e-06 - val_loss: 2.6008e-06 Epoch 419/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 2.1700e-06 - val_loss: 3.8838e-06 Epoch 420/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 4.4710e-06 - val_loss: 2.4320e-06 Epoch 421/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.2794e-05 - val_loss: 6.3769e-06 Epoch 422/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 3.1247e-06 - val_loss: 1.8746e-06 Epoch 423/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 7.1004e-06 - val_loss: 2.6985e-05 Epoch 424/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 5.7805e-06 - val_loss: 2.3848e-06 Epoch 425/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 3.8714e-06 - val_loss: 3.7967e-06 Epoch 426/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 3.9854e-06 - val_loss: 2.9499e-06 Epoch 427/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 8.4056e-06 - val_loss: 4.5954e-05 Epoch 428/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 8.2245e-06 - val_loss: 2.6735e-06 Epoch 429/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 3.6811e-06 - val_loss: 1.3119e-04 Epoch 430/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 2.0221e-05 - val_loss: 1.9291e-06 Epoch 431/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 2.2629e-06 - val_loss: 2.0363e-06 Epoch 432/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 2.3224e-06 - val_loss: 4.2338e-06 Epoch 433/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 3.7307e-06 - val_loss: 8.9199e-06 Epoch 434/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 9.5896e-06 - val_loss: 4.3951e-06 Epoch 435/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.3430e-06 - val_loss: 1.0217e-05 Epoch 436/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 8.0118e-06 - val_loss: 4.1067e-06 Epoch 437/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 7.6434e-06 - val_loss: 5.6278e-06 Epoch 438/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 6.3661e-06 - val_loss: 4.5321e-06 Epoch 439/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 3.9819e-06 - val_loss: 5.2804e-06 Epoch 440/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 3.3630e-06 - val_loss: 6.4379e-06 Epoch 441/1000 3888/3888 [==============================] - 2s 507us/sample - loss: 6.2800e-06 - val_loss: 6.3206e-06 Epoch 442/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 4.5450e-06 - val_loss: 3.4282e-05 Epoch 443/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 1.2518e-05 - val_loss: 1.6223e-06 Epoch 444/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 3.3875e-06 - val_loss: 3.1720e-06 Epoch 445/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 1.6098e-05 - val_loss: 2.2443e-06 Epoch 446/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 2.2748e-06 - val_loss: 2.0723e-05 Epoch 447/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 6.6865e-06 - val_loss: 2.5630e-06 Epoch 448/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 3.0153e-06 - val_loss: 2.3608e-06 Epoch 449/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 5.3289e-06 - val_loss: 3.3566e-06 Epoch 450/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 3.8928e-06 - val_loss: 5.7501e-06 Epoch 451/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 6.3512e-06 - val_loss: 3.6718e-06 Epoch 452/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 9.6717e-06 - val_loss: 2.2501e-06 Epoch 453/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 4.0368e-06 - val_loss: 4.3232e-06 Epoch 454/1000 3888/3888 [==============================] - 2s 533us/sample - loss: 4.1641e-06 - val_loss: 1.0186e-04 Epoch 455/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 1.2537e-05 - val_loss: 2.2803e-06 Epoch 456/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 3.9934e-06 - val_loss: 2.7815e-06 Epoch 457/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 1.0332e-05 - val_loss: 1.6721e-05 Epoch 458/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 3.0069e-06 - val_loss: 1.7623e-06 Epoch 459/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 6.7566e-06 - val_loss: 2.2205e-06 Epoch 460/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.0756e-06 - val_loss: 1.5583e-06 Epoch 461/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 4.6386e-06 - val_loss: 7.1363e-06 Epoch 462/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 8.4558e-06 - val_loss: 2.7535e-06 Epoch 463/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 1.5508e-05 - val_loss: 2.3925e-06 Epoch 464/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 1.8030e-06 - val_loss: 2.4028e-06 Epoch 465/1000 3888/3888 [==============================] - 2s 510us/sample - loss: 2.4461e-06 - val_loss: 3.3196e-06 Epoch 466/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 2.5914e-06 - val_loss: 5.0767e-06 Epoch 467/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 9.7569e-06 - val_loss: 9.4748e-06 Epoch 468/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 4.2160e-06 - val_loss: 1.6647e-06 Epoch 469/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 4.7615e-06 - val_loss: 1.6990e-06 Epoch 470/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 7.4687e-06 - val_loss: 1.5413e-05 Epoch 471/1000 3888/3888 [==============================] - 2s 510us/sample - loss: 8.6830e-06 - val_loss: 1.5755e-06 Epoch 472/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 2.1471e-06 - val_loss: 1.6020e-06 Epoch 473/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 1.2604e-05 - val_loss: 8.9875e-06 Epoch 474/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 3.3254e-06 - val_loss: 4.6769e-06 Epoch 475/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 3.5474e-06 - val_loss: 2.2678e-06 Epoch 476/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 6.1097e-06 - val_loss: 2.5484e-05 Epoch 477/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 5.3429e-06 - val_loss: 2.2617e-06 Epoch 478/1000 3888/3888 [==============================] - 2s 534us/sample - loss: 4.6878e-06 - val_loss: 6.2972e-06 Epoch 479/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 7.5191e-06 - val_loss: 1.8592e-06 Epoch 480/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 2.8433e-06 - val_loss: 3.1215e-06 Epoch 481/1000 3888/3888 [==============================] - 2s 533us/sample - loss: 8.3542e-06 - val_loss: 8.2867e-06 Epoch 482/1000 3888/3888 [==============================] - 2s 507us/sample - loss: 5.5501e-06 - val_loss: 1.5051e-05 Epoch 483/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 4.3921e-06 - val_loss: 2.7111e-05 Epoch 484/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 6.0440e-06 - val_loss: 2.0719e-06 Epoch 485/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 3.7278e-06 - val_loss: 1.3995e-06 Epoch 486/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 8.2209e-06 - val_loss: 4.0150e-06 Epoch 487/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 3.9177e-06 - val_loss: 4.7908e-06 Epoch 488/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 6.6267e-06 - val_loss: 1.0626e-04 Epoch 489/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.0011e-05 - val_loss: 1.6490e-06 Epoch 490/1000 3888/3888 [==============================] - 2s 509us/sample - loss: 2.3579e-06 - val_loss: 2.7779e-06 Epoch 491/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 1.3324e-05 - val_loss: 1.9043e-06 Epoch 492/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 2.3793e-06 - val_loss: 2.4515e-05 Epoch 493/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 4.4677e-06 - val_loss: 1.5414e-06 Epoch 494/1000 3888/3888 [==============================] - 2s 507us/sample - loss: 1.3537e-05 - val_loss: 3.1978e-06 Epoch 495/1000 3888/3888 [==============================] - 2s 507us/sample - loss: 4.0056e-06 - val_loss: 2.3614e-06 Epoch 496/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 2.0211e-06 - val_loss: 1.4953e-06 Epoch 497/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 5.8399e-06 - val_loss: 2.4535e-06 Epoch 498/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 5.3319e-06 - val_loss: 8.2319e-06 Epoch 499/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.3444e-05 - val_loss: 1.4691e-06 Epoch 500/1000 3888/3888 [==============================] - 2s 532us/sample - loss: 1.5649e-06 - val_loss: 3.8898e-06 Epoch 501/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 1.7725e-06 - val_loss: 1.3053e-06 Epoch 502/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 9.5350e-06 - val_loss: 2.9043e-06 Epoch 503/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 3.0681e-06 - val_loss: 3.8133e-06 Epoch 504/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 4.8627e-06 - val_loss: 1.2745e-05 Epoch 505/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 8.1209e-06 - val_loss: 2.9752e-06 Epoch 506/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 2.9745e-06 - val_loss: 4.6810e-06 Epoch 507/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 1.1783e-05 - val_loss: 2.1081e-06 Epoch 508/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 1.9571e-06 - val_loss: 1.6244e-06 Epoch 509/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 5.2568e-06 - val_loss: 1.1290e-04 Epoch 510/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 7.9201e-06 - val_loss: 3.9374e-06 Epoch 511/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 2.5577e-06 - val_loss: 4.4031e-06 Epoch 512/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 8.5399e-06 - val_loss: 4.7891e-06 Epoch 513/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 2.9897e-06 - val_loss: 5.3012e-06 Epoch 514/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.4015e-05 - val_loss: 4.5322e-05 Epoch 515/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 2.8631e-06 - val_loss: 2.2241e-06 Epoch 516/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 2.3498e-06 - val_loss: 2.0121e-06 Epoch 517/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 6.3863e-06 - val_loss: 5.5098e-06 Epoch 518/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 4.0633e-06 - val_loss: 3.0095e-06 Epoch 519/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.9980e-05 - val_loss: 2.1042e-06 Epoch 520/1000 3888/3888 [==============================] - 2s 509us/sample - loss: 1.4994e-06 - val_loss: 1.3542e-06 Epoch 521/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.5451e-06 - val_loss: 1.7695e-06 Epoch 522/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 8.3427e-06 - val_loss: 2.0564e-06 Epoch 523/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 2.7872e-06 - val_loss: 1.4690e-06 Epoch 524/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 2.2669e-06 - val_loss: 3.4363e-06 Epoch 525/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 3.3306e-06 - val_loss: 5.6139e-06 Epoch 526/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 7.5951e-06 - val_loss: 5.0140e-06 Epoch 527/1000 3888/3888 [==============================] - 2s 510us/sample - loss: 2.6646e-06 - val_loss: 5.8321e-06 Epoch 528/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 7.6137e-06 - val_loss: 2.4344e-06 Epoch 529/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 2.2535e-06 - val_loss: 2.2041e-06 Epoch 530/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 1.0164e-05 - val_loss: 9.7516e-05 Epoch 531/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 5.0279e-06 - val_loss: 4.8506e-06 Epoch 532/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 4.5835e-06 - val_loss: 1.8972e-06 Epoch 533/1000 3888/3888 [==============================] - 2s 511us/sample - loss: 7.7115e-06 - val_loss: 1.3270e-05 Epoch 534/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.1392e-05 - val_loss: 1.5546e-05 Epoch 535/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 2.4053e-06 - val_loss: 1.5311e-06 Epoch 536/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 2.5473e-06 - val_loss: 1.1458e-05 Epoch 537/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 3.8128e-06 - val_loss: 2.0542e-06 Epoch 538/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 6.2003e-06 - val_loss: 7.1601e-06 Epoch 539/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 4.9087e-06 - val_loss: 1.1894e-05 Epoch 540/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 7.3835e-06 - val_loss: 1.5244e-06 Epoch 541/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 2.2890e-06 - val_loss: 5.2968e-05 Epoch 542/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 7.3445e-06 - val_loss: 3.4704e-06 Epoch 543/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 5.0404e-06 - val_loss: 7.4623e-06 Epoch 544/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 1.3650e-05 - val_loss: 8.8416e-06 Epoch 545/1000 3888/3888 [==============================] - 2s 533us/sample - loss: 1.8725e-06 - val_loss: 3.1696e-06 Epoch 546/1000 3888/3888 [==============================] - 2s 523us/sample - loss: 3.7403e-06 - val_loss: 2.1714e-05 Epoch 547/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 2.7124e-06 - val_loss: 3.8430e-06 Epoch 548/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 5.3578e-06 - val_loss: 1.6417e-06 Epoch 549/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 3.8552e-06 - val_loss: 5.7091e-06 Epoch 550/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 5.8519e-06 - val_loss: 3.7367e-06 Epoch 551/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 3.6996e-06 - val_loss: 1.0334e-05 Epoch 552/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 8.2127e-06 - val_loss: 5.0946e-06 Epoch 553/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 8.8551e-06 - val_loss: 1.0220e-05 Epoch 554/1000 3888/3888 [==============================] - 2s 522us/sample - loss: 4.1572e-06 - val_loss: 1.3035e-05 Epoch 555/1000 3888/3888 [==============================] - 2s 521us/sample - loss: 4.1485e-06 - val_loss: 9.0559e-06 Epoch 556/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 4.9445e-06 - val_loss: 1.7113e-06 Epoch 557/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 3.7571e-06 - val_loss: 2.9121e-06 Epoch 558/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 4.2609e-06 - val_loss: 3.2241e-06 Epoch 559/1000 3888/3888 [==============================] - 2s 515us/sample - loss: 1.1203e-05 - val_loss: 5.7351e-06 Epoch 560/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 2.1763e-06 - val_loss: 2.3035e-06 Epoch 561/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 3.1797e-06 - val_loss: 4.4746e-06 Epoch 562/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 3.3246e-06 - val_loss: 3.8687e-06 Epoch 563/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 7.1640e-06 - val_loss: 1.8002e-06 Epoch 564/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 3.5687e-06 - val_loss: 1.7234e-06 Epoch 565/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 6.7577e-06 - val_loss: 2.9982e-05 Epoch 566/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 4.3330e-06 - val_loss: 6.5001e-06 Epoch 567/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 4.7838e-06 - val_loss: 1.4490e-06 Epoch 568/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 7.5817e-06 - val_loss: 9.6791e-06 Epoch 569/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 8.5611e-06 - val_loss: 6.9920e-06 Epoch 570/1000 3888/3888 [==============================] - 2s 514us/sample - loss: 3.2533e-06 - val_loss: 3.1448e-06 Epoch 571/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 1.9123e-06 - val_loss: 2.6444e-06 Epoch 572/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.1447e-05 - val_loss: 1.8297e-06 Epoch 573/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 3.1469e-06 - val_loss: 1.6002e-06 Epoch 574/1000 3888/3888 [==============================] - 2s 517us/sample - loss: 1.0359e-05 - val_loss: 9.1355e-06 Epoch 575/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 1.7074e-06 - val_loss: 1.6155e-06 Epoch 576/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 1.7158e-06 - val_loss: 2.4014e-06 Epoch 577/1000 3888/3888 [==============================] - 2s 512us/sample - loss: 3.2391e-06 - val_loss: 4.4465e-06 Epoch 578/1000 3888/3888 [==============================] - 2s 529us/sample - loss: 1.0660e-05 - val_loss: 2.1366e-06 Epoch 579/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 3.9415e-06 - val_loss: 6.3034e-06 Epoch 580/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 3.2084e-06 - val_loss: 5.3064e-06 Epoch 581/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 4.1859e-06 - val_loss: 4.1636e-06 Epoch 582/1000 3888/3888 [==============================] - 2s 524us/sample - loss: 5.3319e-06 - val_loss: 2.0582e-06 Epoch 583/1000 3888/3888 [==============================] - 2s 531us/sample - loss: 6.4532e-06 - val_loss: 4.0206e-06 Epoch 584/1000 3888/3888 [==============================] - 2s 527us/sample - loss: 6.1445e-06 - val_loss: 2.4975e-06 Epoch 585/1000 3888/3888 [==============================] - 2s 513us/sample - loss: 2.9030e-06 - val_loss: 2.0058e-06 Epoch 586/1000 3888/3888 [==============================] - 2s 509us/sample - loss: 7.5439e-06 - val_loss: 4.5810e-06 Epoch 587/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 3.2171e-06 - val_loss: 3.8658e-06 Epoch 588/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 5.4219e-06 - val_loss: 2.3746e-06 Epoch 589/1000 3888/3888 [==============================] - 2s 536us/sample - loss: 4.5144e-06 - val_loss: 2.5607e-06 Epoch 590/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 4.0862e-06 - val_loss: 1.9989e-06 Epoch 591/1000 3888/3888 [==============================] - 2s 525us/sample - loss: 1.1080e-05 - val_loss: 2.2968e-06 Epoch 592/1000 3888/3888 [==============================] - 2s 530us/sample - loss: 2.8358e-06 - val_loss: 7.2369e-06 Epoch 593/1000 3888/3888 [==============================] - 2s 516us/sample - loss: 5.8478e-06 - val_loss: 7.0032e-06 Epoch 594/1000 3888/3888 [==============================] - 2s 518us/sample - loss: 4.4308e-06 - val_loss: 4.2108e-06 Epoch 595/1000 3888/3888 [==============================] - 2s 526us/sample - loss: 3.5390e-06 - val_loss: 2.0974e-06 Epoch 596/1000 3888/3888 [==============================] - 2s 520us/sample - loss: 9.5044e-06 - val_loss: 4.3031e-05 Epoch 597/1000 3888/3888 [==============================] - 2s 536us/sample - loss: 7.6600e-06 - val_loss: 2.7332e-06 Epoch 598/1000 3888/3888 [==============================] - 2s 519us/sample - loss: 1.4903e-06 - val_loss: 1.3590e-06 Epoch 599/1000 3888/3888 [==============================] - 2s 536us/sample - loss: 2.4768e-06 - val_loss: 7.2678e-05 Epoch 600/1000 3888/3888 [==============================] - 2s 528us/sample - loss: 7.0254e-06 - val_loss: 4.8043e-06 Epoch 601/1000 3840/3888 [============================>.] - ETA: 0s - loss: 2.6166e-06Restoring model weights from the end of the best epoch. 3888/3888 [==============================] - 2s 507us/sample - loss: 2.5999e-06 - val_loss: 1.8691e-06 Epoch 00601: early stopping
In [63]:
print(history.history.keys())
print('best value: ', conv_ae.evaluate(X_train, X_train, verbose=0))
pd.DataFrame(history.history).plot(figsize=(8, 5), logy=True)
plt.grid()
dict_keys(['loss', 'val_loss']) best value: 1.3052720110870932e-06
In [64]:
X_reconstructions = conv_ae.predict(X_train)
X_reconstructions = stdscaler.inverse_transform(np.moveaxis(X_reconstructions,3,1).reshape(len(times),len(group)*nl*nc))
calculateerror(X_train_1D.reshape(len(times),len(groups),nl,nc),
X_reconstructions.reshape(len(times),len(groups),nl,nc),
groups,
print_step=0)
max_abs_error: 6.6978759765625 mean_abs_error: 0.015144726812325737
/home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: divide by zero encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until
In [65]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(X_reconstructions.reshape(len(times),len(groups),nl,nc)[100,i,:,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [66]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_train_1D[:,i*nl*nc+4])
ax.flatten()[i].plot(times, X_reconstructions[:,i*nl*nc+4],'--')
ax.flatten()[i].set_title(group)
In [67]:
tf.random.set_seed(42)
np.random.seed(42)
# Need to have validation loss
early_stopping = keras.callbacks.EarlyStopping(monitor='val_loss',
min_delta=0.0,
patience=100,
verbose=2,
restore_best_weights=True)
conv_encoder = keras.models.Sequential([
keras.layers.Reshape([10, 10, 8, 1], input_shape=[10, 10, 8]),
#keras.layers.InputLayer(input_shape=(10, 10, 8)),
keras.layers.Conv3D(64, kernel_size=3, padding="SAME", activation="elu"),
keras.layers.MaxPool3D(pool_size=2),
keras.layers.Conv3D(64, kernel_size=3, padding="SAME", activation="elu"),
keras.layers.MaxPool3D(pool_size=2),
keras.layers.Conv3D(64, kernel_size=3, padding="SAME", activation="elu"),
keras.layers.MaxPool3D(pool_size=2),
keras.layers.Flatten(),
#keras.layers.Dense(64, activation="selu", kernel_initializer="lecun_normal"),
keras.layers.Dense(15)
])
conv_decoder = keras.models.Sequential([
keras.layers.InputLayer(input_shape=(15)),
#keras.layers.Dense(64, activation="selu", kernel_initializer="lecun_normal"),
keras.layers.Dense(64*1*1*1, activation="elu"),
keras.layers.Reshape(target_shape=(1, 1, 1, 64)),
keras.layers.Conv3DTranspose(64, kernel_size=3, strides=2, padding="SAME", activation="elu"),
keras.layers.Conv3DTranspose(64, kernel_size=3, strides=3, padding="SAME", output_padding=[1,1,0], activation="elu"),
keras.layers.Conv3DTranspose(1, kernel_size=3, strides=2, padding="SAME"),
keras.layers.Reshape([10, 10, 8])
])
conv_ae = keras.models.Sequential([conv_encoder, conv_decoder])
conv_ae.compile(loss="mse",
optimizer=keras.optimizers.Nadam(lr=0.0003, beta_1=0.9, beta_2=0.999))
conv_encoder.summary()
conv_decoder.summary()
Model: "sequential_15" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= reshape_2 (Reshape) (None, 10, 10, 8, 1) 0 _________________________________________________________________ conv3d (Conv3D) (None, 10, 10, 8, 64) 1792 _________________________________________________________________ max_pooling3d (MaxPooling3D) (None, 5, 5, 4, 64) 0 _________________________________________________________________ conv3d_1 (Conv3D) (None, 5, 5, 4, 64) 110656 _________________________________________________________________ max_pooling3d_1 (MaxPooling3 (None, 2, 2, 2, 64) 0 _________________________________________________________________ conv3d_2 (Conv3D) (None, 2, 2, 2, 64) 110656 _________________________________________________________________ max_pooling3d_2 (MaxPooling3 (None, 1, 1, 1, 64) 0 _________________________________________________________________ flatten_1 (Flatten) (None, 64) 0 _________________________________________________________________ dense_24 (Dense) (None, 15) 975 ================================================================= Total params: 224,079 Trainable params: 224,079 Non-trainable params: 0 _________________________________________________________________ Model: "sequential_16" _________________________________________________________________ Layer (type) Output Shape Param # ================================================================= dense_25 (Dense) (None, 64) 1024 _________________________________________________________________ reshape_3 (Reshape) (None, 1, 1, 1, 64) 0 _________________________________________________________________ conv3d_transpose (Conv3DTran (None, 2, 2, 2, 64) 110656 _________________________________________________________________ conv3d_transpose_1 (Conv3DTr (None, 5, 5, 4, 64) 110656 _________________________________________________________________ conv3d_transpose_2 (Conv3DTr (None, 10, 10, 8, 1) 1729 _________________________________________________________________ reshape_4 (Reshape) (None, 10, 10, 8) 0 ================================================================= Total params: 224,065 Trainable params: 224,065 Non-trainable params: 0 _________________________________________________________________
In [68]:
history = conv_ae.fit(X_train, X_train,
epochs=1000,
validation_data=(X_train, X_train),
callbacks=[early_stopping])
Train on 3888 samples, validate on 3888 samples Epoch 1/1000 3888/3888 [==============================] - 27s 7ms/sample - loss: 0.0645 - val_loss: 0.0179 Epoch 2/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 0.0133 - val_loss: 0.0085 Epoch 3/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 0.0037 - val_loss: 0.0013 Epoch 4/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 0.0013 - val_loss: 6.0102e-04 Epoch 5/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 7.0133e-04 - val_loss: 4.6277e-04 Epoch 6/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.3439e-04 - val_loss: 2.6882e-04 Epoch 7/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 6.1017e-04 - val_loss: 3.0444e-04 Epoch 8/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.3062e-04 - val_loss: 2.3250e-04 Epoch 9/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.2092e-04 - val_loss: 1.4888e-04 Epoch 10/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.7692e-04 - val_loss: 0.0011 Epoch 11/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.2421e-04 - val_loss: 1.9130e-04 Epoch 12/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0109e-04 - val_loss: 3.9702e-04 Epoch 13/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.8472e-04 - val_loss: 8.7998e-05 Epoch 14/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6230e-04 - val_loss: 3.5600e-04 Epoch 15/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4391e-04 - val_loss: 1.8714e-04 Epoch 16/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.7324e-04 - val_loss: 1.1555e-04 Epoch 17/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9702e-04 - val_loss: 6.2670e-05 Epoch 18/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.3896e-05 - val_loss: 9.0681e-05 Epoch 19/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9174e-04 - val_loss: 6.6589e-05 Epoch 20/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.7725e-05 - val_loss: 5.5764e-05 Epoch 21/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5372e-04 - val_loss: 4.7542e-05 Epoch 22/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 9.9465e-05 - val_loss: 7.8597e-05 Epoch 23/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0755e-04 - val_loss: 4.5444e-05 Epoch 24/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 1.3892e-04 - val_loss: 4.4189e-05 Epoch 25/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 1.1894e-04 - val_loss: 6.5300e-05 Epoch 26/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2847e-04 - val_loss: 4.6664e-04 Epoch 27/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.2728e-05 - val_loss: 7.5831e-05 Epoch 28/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8525e-04 - val_loss: 9.8349e-05 Epoch 29/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0758e-04 - val_loss: 4.4537e-05 Epoch 30/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7128e-04 - val_loss: 1.1983e-04 Epoch 31/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0244e-04 - val_loss: 4.9485e-05 Epoch 32/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.5086e-05 - val_loss: 4.3630e-05 Epoch 33/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.8355e-04 - val_loss: 3.9957e-05 Epoch 34/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.6548e-05 - val_loss: 3.1633e-05 Epoch 35/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0646e-05 - val_loss: 7.3582e-05 Epoch 36/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.4759e-05 - val_loss: 3.8306e-04 Epoch 37/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.6051e-05 - val_loss: 2.8212e-05 Epoch 38/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6551e-04 - val_loss: 2.6479e-05 Epoch 39/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4648e-05 - val_loss: 2.4684e-05 Epoch 40/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.1989e-05 - val_loss: 3.7292e-05 Epoch 41/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0074e-04 - val_loss: 6.6136e-05 Epoch 42/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.2945e-05 - val_loss: 0.0019 Epoch 43/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4815e-04 - val_loss: 3.5411e-04 Epoch 44/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.7046e-05 - val_loss: 1.3403e-04 Epoch 45/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.8387e-05 - val_loss: 1.9411e-05 Epoch 46/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.1307e-05 - val_loss: 2.5702e-05 Epoch 47/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6230e-04 - val_loss: 2.9720e-05 Epoch 48/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.0510e-05 - val_loss: 1.6476e-05 Epoch 49/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1958e-04 - val_loss: 2.2765e-05 Epoch 50/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9401e-05 - val_loss: 1.3560e-04 Epoch 51/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.7963e-05 - val_loss: 8.4380e-05 Epoch 52/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3049e-04 - val_loss: 1.8876e-05 Epoch 53/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.8607e-05 - val_loss: 1.7366e-05 Epoch 54/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.7766e-05 - val_loss: 7.3659e-05 Epoch 55/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9747e-05 - val_loss: 1.3606e-05 Epoch 56/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6305e-05 - val_loss: 2.3468e-05 Epoch 57/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5085e-05 - val_loss: 2.7569e-05 Epoch 58/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.7372e-05 - val_loss: 3.8003e-04 Epoch 59/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.9448e-05 - val_loss: 2.3421e-05 Epoch 60/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8261e-05 - val_loss: 1.8667e-05 Epoch 61/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.9535e-05 - val_loss: 5.7787e-05 Epoch 62/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.4707e-05 - val_loss: 1.2709e-04 Epoch 63/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0013e-05 - val_loss: 1.5185e-05 Epoch 64/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9634e-05 - val_loss: 2.4384e-05 Epoch 65/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.6261e-05 - val_loss: 3.4582e-04 Epoch 66/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.0195e-05 - val_loss: 2.3718e-05 Epoch 67/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7817e-04 - val_loss: 1.6552e-04 Epoch 68/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7320e-05 - val_loss: 1.1497e-05 Epoch 69/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7949e-04 - val_loss: 2.0849e-05 Epoch 70/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6398e-05 - val_loss: 1.0609e-05 Epoch 71/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0537e-05 - val_loss: 1.0104e-05 Epoch 72/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5248e-05 - val_loss: 8.1880e-06 Epoch 73/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8968e-05 - val_loss: 2.1065e-05 Epoch 74/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5661e-05 - val_loss: 1.1680e-05 Epoch 75/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.7603e-05 - val_loss: 4.5737e-05 Epoch 76/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7768e-05 - val_loss: 7.1735e-06 Epoch 77/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.7030e-05 - val_loss: 1.6333e-04 Epoch 78/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.2020e-05 - val_loss: 1.7602e-05 Epoch 79/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5399e-05 - val_loss: 6.1720e-04 Epoch 80/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7780e-04 - val_loss: 1.1774e-05 Epoch 81/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0181e-05 - val_loss: 1.3716e-05 Epoch 82/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0030e-05 - val_loss: 1.4008e-05 Epoch 83/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9977e-05 - val_loss: 7.6291e-06 Epoch 84/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.5692e-05 - val_loss: 9.0191e-06 Epoch 85/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.3797e-05 - val_loss: 1.3189e-05 Epoch 86/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1041e-05 - val_loss: 7.2317e-06 Epoch 87/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5371e-05 - val_loss: 1.1684e-05 Epoch 88/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.9545e-05 - val_loss: 8.6677e-04 Epoch 89/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.8523e-05 - val_loss: 1.1052e-05 Epoch 90/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3515e-05 - val_loss: 8.0050e-06 Epoch 91/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7114e-05 - val_loss: 1.7020e-05 Epoch 92/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.5624e-05 - val_loss: 2.5560e-05 Epoch 93/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3934e-05 - val_loss: 1.1292e-05 Epoch 94/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.4439e-06 - val_loss: 3.1204e-05 Epoch 95/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.4749e-05 - val_loss: 9.1337e-06 Epoch 96/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.4868e-05 - val_loss: 3.6920e-05 Epoch 97/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3189e-05 - val_loss: 3.5045e-05 Epoch 98/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3189e-05 - val_loss: 1.6266e-05 Epoch 99/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0070e-04 - val_loss: 5.8634e-04 Epoch 100/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.5690e-05 - val_loss: 6.6018e-06 Epoch 101/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.0481e-06 - val_loss: 1.5554e-05 Epoch 102/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3833e-05 - val_loss: 6.8025e-06 Epoch 103/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1319e-05 - val_loss: 1.0299e-05 Epoch 104/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4185e-05 - val_loss: 2.4188e-05 Epoch 105/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4697e-05 - val_loss: 7.7796e-06 Epoch 106/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9058e-05 - val_loss: 7.4796e-06 Epoch 107/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 7.8816e-06 - val_loss: 6.0071e-06 Epoch 108/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 4.0847e-05 - val_loss: 6.6178e-06 Epoch 109/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.3560e-05 - val_loss: 1.8455e-05 Epoch 110/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3484e-05 - val_loss: 5.9129e-06 Epoch 111/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5071e-05 - val_loss: 7.2685e-06 Epoch 112/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.9222e-06 - val_loss: 7.1301e-06 Epoch 113/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0648e-05 - val_loss: 7.6133e-06 Epoch 114/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.7882e-05 - val_loss: 4.7872e-05 Epoch 115/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0517e-05 - val_loss: 5.6864e-06 Epoch 116/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.6399e-05 - val_loss: 6.0527e-05 Epoch 117/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6349e-05 - val_loss: 8.2738e-06 Epoch 118/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0886e-05 - val_loss: 1.0745e-05 Epoch 119/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6089e-05 - val_loss: 6.0677e-05 Epoch 120/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8486e-05 - val_loss: 1.2740e-05 Epoch 121/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0335e-05 - val_loss: 2.3188e-04 Epoch 122/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2030e-05 - val_loss: 5.5540e-06 Epoch 123/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8171e-05 - val_loss: 2.8626e-05 Epoch 124/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.9201e-05 - val_loss: 8.6804e-06 Epoch 125/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6662e-05 - val_loss: 1.5689e-05 Epoch 126/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9150e-05 - val_loss: 6.2886e-06 Epoch 127/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.6865e-05 - val_loss: 2.0184e-04 Epoch 128/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3615e-05 - val_loss: 1.4977e-05 Epoch 129/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1761e-05 - val_loss: 5.7443e-06 Epoch 130/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.9165e-06 - val_loss: 3.8631e-05 Epoch 131/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.4477e-05 - val_loss: 1.1585e-05 Epoch 132/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7978e-05 - val_loss: 1.1998e-05 Epoch 133/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0304e-05 - val_loss: 1.5677e-05 Epoch 134/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.0340e-05 - val_loss: 1.2288e-05 Epoch 135/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.1057e-06 - val_loss: 7.3793e-06 Epoch 136/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3524e-05 - val_loss: 5.2595e-06 Epoch 137/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5441e-05 - val_loss: 3.7844e-05 Epoch 138/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3787e-05 - val_loss: 6.7717e-06 Epoch 139/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3705e-05 - val_loss: 1.5190e-05 Epoch 140/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4421e-05 - val_loss: 2.1593e-05 Epoch 141/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.9834e-05 - val_loss: 9.6887e-05 Epoch 142/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 1.1577e-05 - val_loss: 4.2448e-06 Epoch 143/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0103e-05 - val_loss: 1.8027e-05 Epoch 144/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.2469e-06 - val_loss: 9.7967e-06 Epoch 145/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.0995e-05 - val_loss: 4.9933e-06 Epoch 146/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 4.4109e-06 - val_loss: 3.6676e-06 Epoch 147/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.9316e-06 - val_loss: 5.0144e-06 Epoch 148/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5957e-05 - val_loss: 1.7697e-05 Epoch 149/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1266e-05 - val_loss: 1.3042e-05 Epoch 150/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.0068e-05 - val_loss: 1.2905e-04 Epoch 151/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.9922e-06 - val_loss: 4.5736e-06 Epoch 152/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.9870e-06 - val_loss: 1.2870e-05 Epoch 153/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7642e-05 - val_loss: 9.4309e-06 Epoch 154/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.8421e-06 - val_loss: 3.7391e-05 Epoch 155/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.8295e-05 - val_loss: 6.5325e-06 Epoch 156/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.5449e-06 - val_loss: 8.1202e-06 Epoch 157/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.9095e-05 - val_loss: 3.6105e-06 Epoch 158/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.4223e-06 - val_loss: 4.1640e-06 Epoch 159/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.2373e-06 - val_loss: 1.1816e-05 Epoch 160/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.2474e-06 - val_loss: 2.4496e-05 Epoch 161/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7444e-05 - val_loss: 1.0854e-05 Epoch 162/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0934e-05 - val_loss: 2.4965e-05 Epoch 163/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.7011e-05 - val_loss: 4.4553e-06 Epoch 164/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3902e-05 - val_loss: 5.8486e-06 Epoch 165/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0193e-05 - val_loss: 2.5035e-05 Epoch 166/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.3038e-05 - val_loss: 4.5889e-06 Epoch 167/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.3911e-06 - val_loss: 1.1043e-05 Epoch 168/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3721e-05 - val_loss: 6.4711e-06 Epoch 169/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.0874e-06 - val_loss: 1.4516e-05 Epoch 170/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1268e-05 - val_loss: 1.1083e-05 Epoch 171/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8904e-05 - val_loss: 2.6908e-05 Epoch 172/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4403e-05 - val_loss: 8.4103e-06 Epoch 173/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.6878e-06 - val_loss: 1.3387e-05 Epoch 174/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.4919e-05 - val_loss: 5.7586e-06 Epoch 175/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.7270e-06 - val_loss: 3.7061e-06 Epoch 176/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5859e-06 - val_loss: 3.7121e-06 Epoch 177/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4221e-05 - val_loss: 6.4429e-05 Epoch 178/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6768e-05 - val_loss: 4.9492e-06 Epoch 179/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.4355e-06 - val_loss: 1.0123e-05 Epoch 180/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8337e-05 - val_loss: 1.1122e-05 Epoch 181/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.1550e-05 - val_loss: 8.0207e-06 Epoch 182/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 4.9660e-06 - val_loss: 2.8889e-06 Epoch 183/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.0922e-06 - val_loss: 1.5808e-05 Epoch 184/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4978e-05 - val_loss: 2.9725e-06 Epoch 185/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.8162e-05 - val_loss: 4.2424e-06 Epoch 186/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.0940e-06 - val_loss: 3.6263e-06 Epoch 187/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.7531e-06 - val_loss: 7.4437e-06 Epoch 188/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.4290e-06 - val_loss: 5.9724e-06 Epoch 189/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5854e-05 - val_loss: 5.7907e-06 Epoch 190/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3655e-05 - val_loss: 6.7966e-05 Epoch 191/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4272e-05 - val_loss: 5.0592e-06 Epoch 192/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.8572e-06 - val_loss: 3.4861e-06 Epoch 193/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.5218e-05 - val_loss: 3.7197e-06 Epoch 194/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.2828e-06 - val_loss: 4.7497e-06 Epoch 195/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0605e-05 - val_loss: 5.8219e-06 Epoch 196/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.9995e-06 - val_loss: 9.1424e-06 Epoch 197/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.6319e-05 - val_loss: 3.6229e-06 Epoch 198/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.2631e-06 - val_loss: 3.4180e-05 Epoch 199/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6503e-05 - val_loss: 3.8556e-06 Epoch 200/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.5699e-06 - val_loss: 5.1344e-06 Epoch 201/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0567e-05 - val_loss: 4.8763e-06 Epoch 202/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.7282e-06 - val_loss: 5.9419e-06 Epoch 203/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2507e-05 - val_loss: 9.5628e-06 Epoch 204/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.4946e-06 - val_loss: 5.7195e-06 Epoch 205/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4098e-05 - val_loss: 5.1858e-06 Epoch 206/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8403e-05 - val_loss: 2.5995e-05 Epoch 207/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3422e-05 - val_loss: 1.4062e-05 Epoch 208/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.5878e-06 - val_loss: 1.9267e-05 Epoch 209/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.4157e-06 - val_loss: 1.4055e-05 Epoch 210/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3906e-05 - val_loss: 6.0141e-06 Epoch 211/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0534e-05 - val_loss: 1.3446e-04 Epoch 212/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5537e-05 - val_loss: 3.2733e-06 Epoch 213/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.7842e-06 - val_loss: 3.1914e-06 Epoch 214/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6023e-05 - val_loss: 4.4483e-06 Epoch 215/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.0280e-06 - val_loss: 8.7618e-06 Epoch 216/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8114e-05 - val_loss: 1.0879e-05 Epoch 217/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0479e-05 - val_loss: 4.2874e-06 Epoch 218/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.5584e-06 - val_loss: 8.7401e-06 Epoch 219/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8559e-04 - val_loss: 1.6975e-04 Epoch 220/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7733e-05 - val_loss: 3.9709e-06 Epoch 221/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2597e-06 - val_loss: 2.8213e-06 Epoch 222/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5294e-06 - val_loss: 6.0359e-06 Epoch 223/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.1256e-06 - val_loss: 6.8854e-06 Epoch 224/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.0661e-06 - val_loss: 1.6016e-05 Epoch 225/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2161e-05 - val_loss: 9.0677e-06 Epoch 226/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.1705e-06 - val_loss: 2.8227e-06 Epoch 227/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2861e-06 - val_loss: 2.3906e-06 Epoch 228/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.0140e-06 - val_loss: 1.0936e-05 Epoch 229/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1946e-05 - val_loss: 2.7610e-05 Epoch 230/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.9446e-06 - val_loss: 4.4194e-06 Epoch 231/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0146e-05 - val_loss: 1.2693e-05 Epoch 232/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0540e-05 - val_loss: 2.8234e-05 Epoch 233/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5211e-05 - val_loss: 1.3661e-05 Epoch 234/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.6076e-06 - val_loss: 4.1647e-06 Epoch 235/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3647e-05 - val_loss: 2.5881e-06 Epoch 236/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.7987e-05 - val_loss: 5.7978e-04 Epoch 237/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8957e-05 - val_loss: 2.5036e-06 Epoch 238/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6504e-06 - val_loss: 2.8019e-06 Epoch 239/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.5429e-06 - val_loss: 2.0502e-04 Epoch 240/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0832e-05 - val_loss: 4.1547e-06 Epoch 241/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.1662e-06 - val_loss: 4.7692e-06 Epoch 242/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.1077e-06 - val_loss: 5.5972e-05 Epoch 243/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.2010e-06 - val_loss: 7.8137e-06 Epoch 244/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1416e-05 - val_loss: 1.1026e-05 Epoch 245/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.8985e-05 - val_loss: 2.9961e-06 Epoch 246/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.7399e-06 - val_loss: 6.7300e-06 Epoch 247/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.6717e-06 - val_loss: 2.8875e-06 Epoch 248/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5843e-05 - val_loss: 2.1691e-05 Epoch 249/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1638e-06 - val_loss: 6.2934e-06 Epoch 250/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.2259e-05 - val_loss: 1.4772e-05 Epoch 251/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6793e-05 - val_loss: 5.3935e-06 Epoch 252/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0534e-06 - val_loss: 2.0750e-06 Epoch 253/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2952e-06 - val_loss: 7.0304e-06 Epoch 254/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0320e-05 - val_loss: 4.9352e-04 Epoch 255/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6670e-05 - val_loss: 3.5650e-06 Epoch 256/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.9173e-06 - val_loss: 8.9296e-06 Epoch 257/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4036e-05 - val_loss: 8.8567e-05 Epoch 258/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.2342e-06 - val_loss: 4.2321e-06 Epoch 259/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8696e-06 - val_loss: 5.5012e-06 Epoch 260/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.6820e-06 - val_loss: 3.0284e-06 Epoch 261/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7016e-05 - val_loss: 4.4189e-06 Epoch 262/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.8479e-06 - val_loss: 3.1388e-06 Epoch 263/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.4175e-06 - val_loss: 4.1705e-06 Epoch 264/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6407e-05 - val_loss: 8.5656e-06 Epoch 265/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.5397e-06 - val_loss: 7.7751e-06 Epoch 266/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6146e-05 - val_loss: 2.6952e-06 Epoch 267/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.8229e-06 - val_loss: 7.9935e-06 Epoch 268/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9262e-05 - val_loss: 1.9356e-04 Epoch 269/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1172e-05 - val_loss: 2.3722e-05 Epoch 270/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.7688e-06 - val_loss: 4.1298e-06 Epoch 271/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4902e-05 - val_loss: 3.9850e-06 Epoch 272/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.8587e-06 - val_loss: 3.3881e-06 Epoch 273/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0582e-05 - val_loss: 6.6808e-05 Epoch 274/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8769e-05 - val_loss: 2.3674e-06 Epoch 275/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1449e-05 - val_loss: 1.2749e-05 Epoch 276/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.0836e-06 - val_loss: 8.3561e-06 Epoch 277/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.5984e-06 - val_loss: 3.0729e-06 Epoch 278/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7100e-05 - val_loss: 3.3114e-06 Epoch 279/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.0716e-06 - val_loss: 8.2138e-06 Epoch 280/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0234e-05 - val_loss: 7.0866e-06 Epoch 281/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7404e-05 - val_loss: 4.7518e-06 Epoch 282/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.7398e-06 - val_loss: 4.8252e-06 Epoch 283/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5039e-05 - val_loss: 1.1410e-05 Epoch 284/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9639e-06 - val_loss: 2.7529e-06 Epoch 285/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9719e-06 - val_loss: 4.1002e-06 Epoch 286/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4980e-05 - val_loss: 1.7641e-05 Epoch 287/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5673e-05 - val_loss: 2.8693e-06 Epoch 288/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.8242e-06 - val_loss: 2.3793e-05 Epoch 289/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6544e-05 - val_loss: 4.5206e-06 Epoch 290/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6078e-06 - val_loss: 3.1448e-06 Epoch 291/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9461e-06 - val_loss: 2.3433e-06 Epoch 292/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9831e-06 - val_loss: 9.1226e-06 Epoch 293/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4616e-05 - val_loss: 1.4062e-05 Epoch 294/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.9646e-06 - val_loss: 2.5544e-05 Epoch 295/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9192e-05 - val_loss: 1.9452e-05 Epoch 296/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2583e-05 - val_loss: 7.6355e-06 Epoch 297/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.3998e-06 - val_loss: 8.1842e-06 Epoch 298/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6949e-05 - val_loss: 1.1132e-05 Epoch 299/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6054e-05 - val_loss: 5.5370e-06 Epoch 300/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9021e-06 - val_loss: 1.8910e-06 Epoch 301/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0826e-05 - val_loss: 2.6743e-05 Epoch 302/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.9343e-05 - val_loss: 4.4653e-06 Epoch 303/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.6826e-06 - val_loss: 2.6318e-05 Epoch 304/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2288e-05 - val_loss: 5.3681e-06 Epoch 305/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3468e-06 - val_loss: 2.1665e-06 Epoch 306/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.3323e-06 - val_loss: 2.0804e-05 Epoch 307/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0835e-05 - val_loss: 1.8382e-06 Epoch 308/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.2993e-06 - val_loss: 2.4535e-06 Epoch 309/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9745e-05 - val_loss: 1.1990e-05 Epoch 310/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.5036e-06 - val_loss: 2.7715e-06 Epoch 311/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4330e-05 - val_loss: 2.4529e-05 Epoch 312/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.7909e-06 - val_loss: 2.5502e-06 Epoch 313/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8878e-05 - val_loss: 7.1583e-06 Epoch 314/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.2379e-06 - val_loss: 4.8045e-06 Epoch 315/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.2493e-06 - val_loss: 1.3291e-04 Epoch 316/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5995e-05 - val_loss: 7.1144e-06 Epoch 317/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.9824e-06 - val_loss: 5.1017e-06 Epoch 318/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1171e-05 - val_loss: 8.3160e-06 Epoch 319/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.6706e-06 - val_loss: 1.2412e-05 Epoch 320/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0677e-05 - val_loss: 2.9676e-06 Epoch 321/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.2859e-06 - val_loss: 1.1344e-05 Epoch 322/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.0877e-06 - val_loss: 1.5857e-05 Epoch 323/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.8649e-06 - val_loss: 2.0032e-05 Epoch 324/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5161e-05 - val_loss: 1.2625e-05 Epoch 325/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.1235e-06 - val_loss: 3.7669e-06 Epoch 326/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7525e-05 - val_loss: 4.0101e-06 Epoch 327/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.0887e-06 - val_loss: 3.5193e-06 Epoch 328/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6846e-05 - val_loss: 2.1912e-05 Epoch 329/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.8894e-06 - val_loss: 2.7197e-06 Epoch 330/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7503e-05 - val_loss: 2.9765e-06 Epoch 331/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.7761e-06 - val_loss: 7.9608e-05 Epoch 332/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5092e-05 - val_loss: 2.4658e-06 Epoch 333/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.2098e-06 - val_loss: 3.8985e-06 Epoch 334/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4489e-05 - val_loss: 3.0874e-06 Epoch 335/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.7718e-06 - val_loss: 3.2098e-04 Epoch 336/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7231e-05 - val_loss: 1.7458e-06 Epoch 337/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.1549e-06 - val_loss: 6.0096e-06 Epoch 338/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.5611e-06 - val_loss: 2.7532e-05 Epoch 339/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.3851e-06 - val_loss: 6.8447e-06 Epoch 340/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3828e-05 - val_loss: 1.3944e-05 Epoch 341/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7152e-05 - val_loss: 1.1251e-05 Epoch 342/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8053e-06 - val_loss: 4.2677e-06 Epoch 343/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4634e-06 - val_loss: 2.4049e-06 Epoch 344/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7995e-05 - val_loss: 3.1839e-06 Epoch 345/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.5926e-06 - val_loss: 3.5819e-06 Epoch 346/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.2068e-06 - val_loss: 6.2199e-06 Epoch 347/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.3199e-05 - val_loss: 6.5630e-06 Epoch 348/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9872e-06 - val_loss: 6.5838e-06 Epoch 349/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4526e-06 - val_loss: 9.1056e-06 Epoch 350/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0226e-05 - val_loss: 2.1736e-05 Epoch 351/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6254e-05 - val_loss: 1.6571e-06 Epoch 352/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.8570e-06 - val_loss: 5.6372e-06 Epoch 353/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9258e-05 - val_loss: 9.8708e-06 Epoch 354/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4799e-06 - val_loss: 3.0529e-06 Epoch 355/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8970e-05 - val_loss: 9.7103e-05 Epoch 356/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0765e-05 - val_loss: 5.2303e-06 Epoch 357/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.3642e-06 - val_loss: 3.9316e-06 Epoch 358/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7696e-05 - val_loss: 1.2809e-05 Epoch 359/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.5582e-06 - val_loss: 1.6188e-05 Epoch 360/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.6199e-06 - val_loss: 3.3366e-06 Epoch 361/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5320e-05 - val_loss: 5.3130e-05 Epoch 362/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3166e-05 - val_loss: 3.8429e-05 Epoch 363/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1300e-05 - val_loss: 4.0415e-06 Epoch 364/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.8959e-06 - val_loss: 3.0321e-06 Epoch 365/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9981e-06 - val_loss: 7.8480e-06 Epoch 366/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2399e-05 - val_loss: 6.1186e-06 Epoch 367/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8772e-05 - val_loss: 7.9795e-06 Epoch 368/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.8795e-06 - val_loss: 4.2210e-06 Epoch 369/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.7394e-06 - val_loss: 1.3257e-05 Epoch 370/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1198e-05 - val_loss: 7.8698e-06 Epoch 371/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4774e-06 - val_loss: 1.8981e-06 Epoch 372/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4576e-06 - val_loss: 1.6571e-06 Epoch 373/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.3515e-06 - val_loss: 1.5787e-06 Epoch 374/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.3402e-06 - val_loss: 5.0556e-06 Epoch 375/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.4162e-06 - val_loss: 1.2430e-05 Epoch 376/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.9921e-06 - val_loss: 2.1811e-06 Epoch 377/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.7148e-06 - val_loss: 2.3975e-06 Epoch 378/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.8714e-06 - val_loss: 1.9462e-05 Epoch 379/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0419e-05 - val_loss: 1.7294e-04 Epoch 380/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6479e-05 - val_loss: 5.5537e-06 Epoch 381/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7851e-06 - val_loss: 3.7647e-06 Epoch 382/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0539e-05 - val_loss: 2.5294e-06 Epoch 383/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2382e-05 - val_loss: 1.6727e-05 Epoch 384/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5628e-06 - val_loss: 1.6156e-06 Epoch 385/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4070e-06 - val_loss: 2.8558e-05 Epoch 386/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.0750e-06 - val_loss: 5.0933e-06 Epoch 387/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7281e-05 - val_loss: 3.0829e-05 Epoch 388/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1629e-05 - val_loss: 2.5182e-06 Epoch 389/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3222e-06 - val_loss: 2.3697e-06 Epoch 390/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2518e-06 - val_loss: 4.3680e-06 Epoch 391/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7339e-05 - val_loss: 3.7992e-06 Epoch 392/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1709e-05 - val_loss: 1.5287e-05 Epoch 393/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7495e-05 - val_loss: 3.5470e-06 Epoch 394/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.3256e-06 - val_loss: 1.8248e-06 Epoch 395/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2331e-05 - val_loss: 1.4636e-05 Epoch 396/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.2546e-05 - val_loss: 1.0844e-04 Epoch 397/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.1907e-06 - val_loss: 2.0723e-06 Epoch 398/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5610e-06 - val_loss: 1.4328e-06 Epoch 399/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5477e-06 - val_loss: 2.1586e-06 Epoch 400/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8732e-06 - val_loss: 2.8425e-06 Epoch 401/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.3002e-06 - val_loss: 4.4867e-06 Epoch 402/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0430e-05 - val_loss: 7.6091e-06 Epoch 403/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.9386e-06 - val_loss: 1.6681e-05 Epoch 404/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7116e-05 - val_loss: 3.2354e-06 Epoch 405/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1029e-05 - val_loss: 2.9025e-06 Epoch 406/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.8388e-05 - val_loss: 7.9700e-06 Epoch 407/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7480e-06 - val_loss: 3.6721e-06 Epoch 408/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2930e-06 - val_loss: 1.2327e-06 Epoch 409/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5372e-06 - val_loss: 9.5345e-06 Epoch 410/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2082e-05 - val_loss: 1.7983e-06 Epoch 411/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9837e-06 - val_loss: 2.2540e-06 Epoch 412/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9781e-06 - val_loss: 8.1043e-06 Epoch 413/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6810e-05 - val_loss: 1.7045e-06 Epoch 414/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1446e-05 - val_loss: 2.9517e-06 Epoch 415/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9160e-06 - val_loss: 1.1618e-05 Epoch 416/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0410e-06 - val_loss: 2.7848e-06 Epoch 417/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0668e-05 - val_loss: 1.6369e-05 Epoch 418/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.2180e-06 - val_loss: 4.3633e-06 Epoch 419/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1093e-06 - val_loss: 4.9985e-06 Epoch 420/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4571e-05 - val_loss: 2.1615e-06 Epoch 421/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.5644e-06 - val_loss: 3.4162e-06 Epoch 422/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1330e-05 - val_loss: 1.8454e-06 Epoch 423/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9444e-06 - val_loss: 9.8210e-06 Epoch 424/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.6361e-06 - val_loss: 3.3164e-06 Epoch 425/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.5957e-06 - val_loss: 2.8945e-05 Epoch 426/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0282e-05 - val_loss: 1.7842e-05 Epoch 427/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6159e-06 - val_loss: 2.3305e-05 Epoch 428/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8766e-05 - val_loss: 1.2350e-04 Epoch 429/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.0834e-06 - val_loss: 5.1694e-05 Epoch 430/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0295e-05 - val_loss: 1.3140e-06 Epoch 431/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.9117e-06 - val_loss: 3.7196e-05 Epoch 432/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2866e-05 - val_loss: 2.3805e-06 Epoch 433/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1379e-06 - val_loss: 2.6302e-06 Epoch 434/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5981e-06 - val_loss: 1.4342e-05 Epoch 435/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6523e-05 - val_loss: 2.7720e-06 Epoch 436/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3803e-06 - val_loss: 2.5773e-06 Epoch 437/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.1968e-06 - val_loss: 1.0434e-05 Epoch 438/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3427e-05 - val_loss: 1.0898e-05 Epoch 439/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.7490e-06 - val_loss: 2.0043e-05 Epoch 440/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1962e-05 - val_loss: 6.5810e-06 Epoch 441/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7905e-06 - val_loss: 1.7469e-06 Epoch 442/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5582e-06 - val_loss: 2.1498e-05 Epoch 443/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8829e-05 - val_loss: 2.5506e-06 Epoch 444/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.4901e-06 - val_loss: 1.5371e-06 Epoch 445/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6763e-05 - val_loss: 4.0585e-06 Epoch 446/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9339e-06 - val_loss: 2.9940e-06 Epoch 447/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9427e-06 - val_loss: 9.7076e-06 Epoch 448/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.6872e-06 - val_loss: 4.2538e-06 Epoch 449/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4074e-06 - val_loss: 1.6469e-06 Epoch 450/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3592e-05 - val_loss: 5.6535e-06 Epoch 451/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.5212e-06 - val_loss: 1.8185e-05 Epoch 452/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8466e-05 - val_loss: 5.6691e-05 Epoch 453/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.7948e-06 - val_loss: 2.2572e-06 Epoch 454/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5620e-06 - val_loss: 5.9396e-05 Epoch 455/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8740e-05 - val_loss: 3.2102e-06 Epoch 456/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.8863e-06 - val_loss: 3.0595e-05 Epoch 457/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.3089e-05 - val_loss: 4.4024e-06 Epoch 458/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6096e-06 - val_loss: 1.3994e-06 Epoch 459/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2700e-06 - val_loss: 1.3073e-06 Epoch 460/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8707e-06 - val_loss: 1.9146e-06 Epoch 461/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.7898e-06 - val_loss: 4.9201e-06 Epoch 462/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.9175e-06 - val_loss: 1.1871e-06 Epoch 463/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7909e-05 - val_loss: 3.0527e-06 Epoch 464/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.1629e-06 - val_loss: 2.2708e-06 Epoch 465/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5498e-06 - val_loss: 3.2730e-06 Epoch 466/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3103e-05 - val_loss: 1.8674e-05 Epoch 467/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.6882e-06 - val_loss: 1.5706e-06 Epoch 468/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2193e-05 - val_loss: 2.8550e-06 Epoch 469/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.9752e-06 - val_loss: 2.3210e-06 Epoch 470/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.9457e-06 - val_loss: 4.0313e-05 Epoch 471/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0220e-05 - val_loss: 2.6629e-06 Epoch 472/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1160e-06 - val_loss: 2.7336e-06 Epoch 473/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2549e-05 - val_loss: 6.9065e-06 Epoch 474/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7380e-05 - val_loss: 2.7181e-05 Epoch 475/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1516e-06 - val_loss: 1.9019e-06 Epoch 476/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2201e-06 - val_loss: 8.4208e-06 Epoch 477/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6630e-05 - val_loss: 3.5877e-05 Epoch 478/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.3373e-06 - val_loss: 6.2671e-06 Epoch 479/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2239e-06 - val_loss: 1.1050e-06 Epoch 480/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.2317e-06 - val_loss: 3.7757e-06 Epoch 481/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1544e-05 - val_loss: 9.3099e-05 Epoch 482/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.2510e-06 - val_loss: 2.6037e-06 Epoch 483/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.9229e-06 - val_loss: 2.4616e-06 Epoch 484/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.7306e-06 - val_loss: 2.3002e-06 Epoch 485/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2941e-05 - val_loss: 1.6621e-06 Epoch 486/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5915e-06 - val_loss: 1.8730e-06 Epoch 487/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3155e-06 - val_loss: 2.6492e-06 Epoch 488/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.4967e-06 - val_loss: 1.4532e-05 Epoch 489/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0199e-05 - val_loss: 6.8134e-06 Epoch 490/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.5593e-06 - val_loss: 2.7726e-06 Epoch 491/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3992e-05 - val_loss: 4.2188e-06 Epoch 492/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8455e-06 - val_loss: 2.5779e-06 Epoch 493/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7420e-06 - val_loss: 7.9425e-06 Epoch 494/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5136e-05 - val_loss: 3.6791e-06 Epoch 495/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9405e-06 - val_loss: 3.0632e-06 Epoch 496/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.2946e-06 - val_loss: 1.3207e-06 Epoch 497/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.7566e-06 - val_loss: 2.6618e-06 Epoch 498/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.1508e-06 - val_loss: 1.0827e-05 Epoch 499/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8777e-05 - val_loss: 2.9026e-06 Epoch 500/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6672e-06 - val_loss: 1.7816e-06 Epoch 501/1000 3888/3888 [==============================] - 25s 7ms/sample - loss: 1.0302e-05 - val_loss: 4.9677e-06 Epoch 502/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0963e-05 - val_loss: 4.2461e-06 Epoch 503/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.1117e-06 - val_loss: 8.4581e-06 Epoch 504/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.7325e-06 - val_loss: 8.4431e-06 Epoch 505/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0685e-05 - val_loss: 3.4090e-06 Epoch 506/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6380e-06 - val_loss: 4.6476e-06 Epoch 507/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.0606e-06 - val_loss: 1.4625e-06 Epoch 508/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0591e-06 - val_loss: 6.3639e-06 Epoch 509/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1879e-05 - val_loss: 3.5000e-04 Epoch 510/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1970e-05 - val_loss: 2.4689e-06 Epoch 511/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7259e-06 - val_loss: 1.4516e-06 Epoch 512/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.9112e-06 - val_loss: 2.4964e-05 Epoch 513/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3963e-05 - val_loss: 1.5810e-06 Epoch 514/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7692e-05 - val_loss: 3.6818e-05 Epoch 515/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.1967e-06 - val_loss: 1.5123e-06 Epoch 516/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.1382e-06 - val_loss: 4.2376e-06 Epoch 517/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.3197e-06 - val_loss: 2.6531e-06 Epoch 518/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5247e-06 - val_loss: 4.4808e-06 Epoch 519/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.4625e-06 - val_loss: 8.5993e-06 Epoch 520/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.0658e-06 - val_loss: 5.5232e-06 Epoch 521/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0766e-05 - val_loss: 3.2520e-06 Epoch 522/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9603e-06 - val_loss: 2.1044e-06 Epoch 523/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.2425e-06 - val_loss: 5.5724e-06 Epoch 524/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1037e-05 - val_loss: 1.6748e-06 Epoch 525/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1982e-06 - val_loss: 2.3041e-05 Epoch 526/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1935e-05 - val_loss: 2.1598e-06 Epoch 527/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7640e-06 - val_loss: 6.9504e-06 Epoch 528/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5524e-05 - val_loss: 1.6832e-05 Epoch 529/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.8794e-06 - val_loss: 2.7627e-06 Epoch 530/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.1826e-06 - val_loss: 5.9030e-05 Epoch 531/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5602e-05 - val_loss: 0.0013 Epoch 532/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7204e-05 - val_loss: 1.9543e-06 Epoch 533/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9748e-06 - val_loss: 2.5636e-06 Epoch 534/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.4485e-06 - val_loss: 8.0238e-06 Epoch 535/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0148e-06 - val_loss: 1.0977e-06 Epoch 536/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8297e-06 - val_loss: 3.1114e-06 Epoch 537/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7279e-05 - val_loss: 3.4108e-06 Epoch 538/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3361e-06 - val_loss: 3.7248e-06 Epoch 539/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.8194e-06 - val_loss: 4.5008e-06 Epoch 540/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.4833e-06 - val_loss: 1.6702e-06 Epoch 541/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9758e-06 - val_loss: 1.7727e-04 Epoch 542/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6670e-05 - val_loss: 3.5155e-06 Epoch 543/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1810e-06 - val_loss: 2.6525e-06 Epoch 544/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6608e-06 - val_loss: 1.9859e-06 Epoch 545/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.5788e-06 - val_loss: 1.3038e-04 Epoch 546/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1779e-05 - val_loss: 1.6578e-05 Epoch 547/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.9256e-06 - val_loss: 2.2667e-05 Epoch 548/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.6737e-06 - val_loss: 3.6928e-06 Epoch 549/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.2907e-06 - val_loss: 4.1552e-05 Epoch 550/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.8646e-06 - val_loss: 7.4832e-06 Epoch 551/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9641e-06 - val_loss: 5.9681e-05 Epoch 552/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2696e-05 - val_loss: 1.5658e-06 Epoch 553/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6605e-05 - val_loss: 2.3973e-05 Epoch 554/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.5548e-06 - val_loss: 1.6522e-05 Epoch 555/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6798e-06 - val_loss: 3.7012e-06 Epoch 556/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.6500e-06 - val_loss: 1.1675e-06 Epoch 557/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6407e-05 - val_loss: 1.2810e-05 Epoch 558/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5565e-06 - val_loss: 2.2760e-06 Epoch 559/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1756e-05 - val_loss: 1.2330e-05 Epoch 560/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.3257e-06 - val_loss: 2.0348e-06 Epoch 561/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.8081e-06 - val_loss: 1.9160e-06 Epoch 562/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1537e-05 - val_loss: 4.3128e-06 Epoch 563/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.6935e-06 - val_loss: 2.2248e-06 Epoch 564/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5494e-05 - val_loss: 1.4273e-06 Epoch 565/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9940e-05 - val_loss: 4.6616e-06 Epoch 566/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4349e-06 - val_loss: 2.0059e-06 Epoch 567/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.0320e-06 - val_loss: 2.2421e-06 Epoch 568/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5344e-05 - val_loss: 5.3847e-06 Epoch 569/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2267e-06 - val_loss: 1.6299e-06 Epoch 570/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4201e-05 - val_loss: 2.1310e-06 Epoch 571/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6596e-06 - val_loss: 3.2991e-06 Epoch 572/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6744e-05 - val_loss: 1.8938e-06 Epoch 573/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5270e-06 - val_loss: 1.1681e-06 Epoch 574/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1897e-05 - val_loss: 2.5746e-05 Epoch 575/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3719e-06 - val_loss: 2.2581e-06 Epoch 576/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0957e-06 - val_loss: 2.2371e-06 Epoch 577/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9411e-06 - val_loss: 1.8941e-06 Epoch 578/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7142e-05 - val_loss: 1.8515e-06 Epoch 579/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.8793e-06 - val_loss: 6.5662e-06 Epoch 580/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8135e-06 - val_loss: 2.0197e-06 Epoch 581/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6701e-06 - val_loss: 2.8080e-06 Epoch 582/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.5456e-06 - val_loss: 3.4867e-06 Epoch 583/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6729e-05 - val_loss: 2.6111e-05 Epoch 584/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0804e-05 - val_loss: 2.4409e-06 Epoch 585/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7569e-06 - val_loss: 2.7870e-06 Epoch 586/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3594e-06 - val_loss: 8.9507e-06 Epoch 587/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.7586e-06 - val_loss: 1.9234e-05 Epoch 588/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9098e-05 - val_loss: 4.0467e-06 Epoch 589/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1890e-06 - val_loss: 1.8381e-06 Epoch 590/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.4156e-06 - val_loss: 2.6882e-06 Epoch 591/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.2541e-06 - val_loss: 1.3074e-05 Epoch 592/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.6957e-06 - val_loss: 3.3823e-06 Epoch 593/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2607e-05 - val_loss: 4.8906e-06 Epoch 594/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0421e-05 - val_loss: 1.6661e-05 Epoch 595/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1893e-06 - val_loss: 2.2540e-06 Epoch 596/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.6894e-06 - val_loss: 5.1744e-05 Epoch 597/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9054e-05 - val_loss: 5.1232e-06 Epoch 598/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6281e-06 - val_loss: 3.9606e-06 Epoch 599/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3483e-06 - val_loss: 1.3690e-05 Epoch 600/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4790e-05 - val_loss: 2.4143e-06 Epoch 601/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.3944e-06 - val_loss: 3.2145e-06 Epoch 602/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6352e-05 - val_loss: 3.4554e-05 Epoch 603/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1401e-05 - val_loss: 1.5602e-06 Epoch 604/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8149e-06 - val_loss: 6.9155e-06 Epoch 605/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.3295e-06 - val_loss: 8.0885e-06 Epoch 606/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9363e-05 - val_loss: 2.0667e-05 Epoch 607/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.8090e-06 - val_loss: 3.8826e-06 Epoch 608/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9203e-06 - val_loss: 2.7585e-06 Epoch 609/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4750e-06 - val_loss: 2.9368e-06 Epoch 610/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8326e-05 - val_loss: 3.8102e-06 Epoch 611/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.4699e-06 - val_loss: 3.3834e-06 Epoch 612/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.3672e-06 - val_loss: 2.0387e-06 Epoch 613/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7659e-05 - val_loss: 9.2670e-05 Epoch 614/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.9818e-06 - val_loss: 3.6425e-06 Epoch 615/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.5519e-06 - val_loss: 2.3307e-06 Epoch 616/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.3608e-06 - val_loss: 5.8237e-06 Epoch 617/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9086e-05 - val_loss: 1.2601e-05 Epoch 618/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.3313e-06 - val_loss: 2.3233e-06 Epoch 619/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.4432e-06 - val_loss: 3.7451e-06 Epoch 620/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2723e-06 - val_loss: 1.7335e-06 Epoch 621/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.9868e-06 - val_loss: 6.9475e-06 Epoch 622/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0114e-05 - val_loss: 9.3069e-06 Epoch 623/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8124e-05 - val_loss: 1.7672e-06 Epoch 624/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3989e-06 - val_loss: 1.1625e-06 Epoch 625/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.6438e-06 - val_loss: 1.7478e-06 Epoch 626/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7283e-05 - val_loss: 1.4054e-05 Epoch 627/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.7362e-06 - val_loss: 4.6295e-06 Epoch 628/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7601e-05 - val_loss: 3.3281e-06 Epoch 629/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5111e-06 - val_loss: 1.5195e-06 Epoch 630/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0133e-06 - val_loss: 3.9153e-06 Epoch 631/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.5357e-06 - val_loss: 2.0025e-06 Epoch 632/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.1802e-05 - val_loss: 6.7810e-06 Epoch 633/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5305e-06 - val_loss: 1.2397e-06 Epoch 634/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.7975e-06 - val_loss: 1.0428e-06 Epoch 635/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3612e-06 - val_loss: 2.1498e-06 Epoch 636/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.4041e-06 - val_loss: 1.5012e-06 Epoch 637/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6371e-05 - val_loss: 2.4546e-06 Epoch 638/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.7082e-06 - val_loss: 2.3361e-06 Epoch 639/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3605e-05 - val_loss: 1.2134e-04 Epoch 640/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.8022e-06 - val_loss: 2.6426e-06 Epoch 641/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3676e-06 - val_loss: 1.5212e-06 Epoch 642/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9690e-06 - val_loss: 6.8162e-06 Epoch 643/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5145e-05 - val_loss: 1.6881e-06 Epoch 644/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.0873e-06 - val_loss: 2.1053e-06 Epoch 645/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9724e-06 - val_loss: 2.2792e-06 Epoch 646/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.1464e-06 - val_loss: 5.7513e-06 Epoch 647/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5772e-05 - val_loss: 2.4785e-06 Epoch 648/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4268e-06 - val_loss: 2.7976e-06 Epoch 649/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6196e-06 - val_loss: 2.4775e-06 Epoch 650/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.7566e-06 - val_loss: 1.3547e-06 Epoch 651/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.0992e-06 - val_loss: 6.2714e-05 Epoch 652/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.1923e-06 - val_loss: 7.2175e-06 Epoch 653/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.9561e-06 - val_loss: 1.9828e-06 Epoch 654/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5787e-05 - val_loss: 2.0099e-06 Epoch 655/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0915e-06 - val_loss: 2.1493e-06 Epoch 656/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9732e-06 - val_loss: 9.0273e-07 Epoch 657/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.7884e-06 - val_loss: 3.6888e-06 Epoch 658/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.9629e-06 - val_loss: 1.7047e-06 Epoch 659/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.2508e-06 - val_loss: 6.4187e-06 Epoch 660/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4510e-05 - val_loss: 9.3093e-05 Epoch 661/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.0446e-06 - val_loss: 2.6692e-06 Epoch 662/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5187e-06 - val_loss: 2.9071e-06 Epoch 663/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.7358e-06 - val_loss: 1.4749e-06 Epoch 664/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2927e-06 - val_loss: 2.4128e-06 Epoch 665/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.3654e-06 - val_loss: 3.9152e-06 Epoch 666/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2475e-05 - val_loss: 6.8375e-06 Epoch 667/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.8612e-06 - val_loss: 2.2017e-06 Epoch 668/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.6217e-06 - val_loss: 5.8312e-05 Epoch 669/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.6991e-06 - val_loss: 2.1325e-06 Epoch 670/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.6007e-06 - val_loss: 1.4804e-05 Epoch 671/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.1047e-06 - val_loss: 1.4811e-06 Epoch 672/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.0535e-06 - val_loss: 3.1895e-06 Epoch 673/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0231e-05 - val_loss: 2.0922e-06 Epoch 674/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.7551e-06 - val_loss: 4.2293e-05 Epoch 675/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.0426e-06 - val_loss: 6.0160e-06 Epoch 676/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.3599e-06 - val_loss: 1.2902e-06 Epoch 677/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.8338e-06 - val_loss: 1.0942e-06 Epoch 678/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.4045e-05 - val_loss: 3.4290e-06 Epoch 679/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.4639e-06 - val_loss: 3.8250e-06 Epoch 680/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.9980e-06 - val_loss: 4.8498e-06 Epoch 681/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.9393e-06 - val_loss: 4.3163e-06 Epoch 682/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.3280e-06 - val_loss: 1.7860e-06 Epoch 683/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6617e-05 - val_loss: 3.0099e-05 Epoch 684/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6069e-06 - val_loss: 3.3856e-06 Epoch 685/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0469e-05 - val_loss: 2.4201e-06 Epoch 686/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5942e-06 - val_loss: 1.5469e-06 Epoch 687/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4221e-06 - val_loss: 1.6682e-06 Epoch 688/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.9975e-06 - val_loss: 1.2484e-05 Epoch 689/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.3706e-05 - val_loss: 5.4635e-06 Epoch 690/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.1193e-06 - val_loss: 3.9881e-06 Epoch 691/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.9073e-06 - val_loss: 2.4884e-06 Epoch 692/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0081e-05 - val_loss: 6.6150e-06 Epoch 693/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9668e-05 - val_loss: 2.3831e-06 Epoch 694/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.6483e-06 - val_loss: 1.9681e-06 Epoch 695/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5607e-06 - val_loss: 2.0258e-05 Epoch 696/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2006e-05 - val_loss: 2.4010e-05 Epoch 697/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.0877e-06 - val_loss: 1.1488e-06 Epoch 698/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.7557e-06 - val_loss: 1.8847e-06 Epoch 699/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.5033e-06 - val_loss: 5.9241e-06 Epoch 700/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.3913e-06 - val_loss: 5.7067e-05 Epoch 701/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0496e-05 - val_loss: 2.0991e-06 Epoch 702/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.8951e-06 - val_loss: 1.1018e-05 Epoch 703/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.8410e-06 - val_loss: 1.1916e-05 Epoch 704/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.3333e-06 - val_loss: 1.2308e-05 Epoch 705/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.0848e-06 - val_loss: 4.4458e-06 Epoch 706/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.5756e-06 - val_loss: 1.2734e-06 Epoch 707/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.6885e-06 - val_loss: 2.1342e-05 Epoch 708/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8995e-05 - val_loss: 1.7240e-06 Epoch 709/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.1983e-06 - val_loss: 5.0669e-06 Epoch 710/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9963e-06 - val_loss: 1.1463e-06 Epoch 711/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.4037e-05 - val_loss: 2.8233e-06 Epoch 712/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9458e-06 - val_loss: 1.7376e-06 Epoch 713/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.2944e-06 - val_loss: 2.6379e-06 Epoch 714/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2434e-06 - val_loss: 1.2183e-06 Epoch 715/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2757e-06 - val_loss: 1.3196e-05 Epoch 716/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2646e-05 - val_loss: 1.0804e-05 Epoch 717/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.5773e-06 - val_loss: 2.1277e-06 Epoch 718/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1450e-05 - val_loss: 1.5117e-06 Epoch 719/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8286e-06 - val_loss: 2.1248e-06 Epoch 720/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.9266e-06 - val_loss: 1.5344e-05 Epoch 721/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.4837e-06 - val_loss: 1.3613e-06 Epoch 722/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.3368e-06 - val_loss: 1.0740e-05 Epoch 723/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.7202e-06 - val_loss: 4.9885e-06 Epoch 724/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 9.8470e-06 - val_loss: 1.7056e-05 Epoch 725/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.2041e-05 - val_loss: 5.5972e-06 Epoch 726/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.9560e-06 - val_loss: 1.6328e-06 Epoch 727/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.8258e-06 - val_loss: 3.8789e-06 Epoch 728/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.1156e-05 - val_loss: 2.1728e-06 Epoch 729/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2720e-05 - val_loss: 4.2272e-06 Epoch 730/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.8056e-06 - val_loss: 8.4858e-06 Epoch 731/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1093e-06 - val_loss: 1.9013e-06 Epoch 732/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 7.7393e-06 - val_loss: 8.4994e-06 Epoch 733/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.7974e-06 - val_loss: 3.6759e-06 Epoch 734/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.1015e-06 - val_loss: 2.7462e-05 Epoch 735/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0916e-05 - val_loss: 1.4867e-06 Epoch 736/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0871e-06 - val_loss: 2.1985e-06 Epoch 737/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.0223e-05 - val_loss: 5.1013e-06 Epoch 738/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0697e-06 - val_loss: 1.2254e-06 Epoch 739/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.5044e-06 - val_loss: 5.0729e-06 Epoch 740/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.2617e-05 - val_loss: 6.2119e-06 Epoch 741/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2197e-06 - val_loss: 1.0290e-05 Epoch 742/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5227e-05 - val_loss: 4.5622e-06 Epoch 743/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.0746e-06 - val_loss: 2.5829e-05 Epoch 744/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.6144e-06 - val_loss: 2.1563e-06 Epoch 745/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.6209e-06 - val_loss: 8.2831e-06 Epoch 746/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4346e-05 - val_loss: 2.6101e-06 Epoch 747/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5366e-06 - val_loss: 1.5143e-06 Epoch 748/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 3.2409e-06 - val_loss: 5.0632e-06 Epoch 749/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 8.2085e-06 - val_loss: 1.0033e-04 Epoch 750/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 5.4396e-06 - val_loss: 3.2083e-06 Epoch 751/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.4178e-06 - val_loss: 1.3873e-05 Epoch 752/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 1.5007e-05 - val_loss: 1.1392e-06 Epoch 753/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.1566e-06 - val_loss: 8.0894e-06 Epoch 754/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 6.3194e-06 - val_loss: 2.9919e-05 Epoch 755/1000 3888/3888 [==============================] - 25s 6ms/sample - loss: 4.2546e-05 - val_loss: 5.6485e-06 Epoch 756/1000 3872/3888 [============================>.] - ETA: 0s - loss: 2.1037e-06Restoring model weights from the end of the best epoch. 3888/3888 [==============================] - 25s 6ms/sample - loss: 2.0990e-06 - val_loss: 1.0192e-06 Epoch 00756: early stopping
In [69]:
print(history.history.keys())
print('best value: ', conv_ae.evaluate(X_train, X_train, verbose=0))
pd.DataFrame(history.history).plot(figsize=(8, 5), logy=True)
plt.grid()
dict_keys(['loss', 'val_loss']) best value: 9.027328154536509e-07
In [70]:
X_reconstructions = conv_ae.predict(X_train)
X_reconstructions = stdscaler.inverse_transform(np.moveaxis(X_reconstructions,3,1).reshape(len(times),len(group)*nl*nc))
calculateerror(X_train_1D.reshape(len(times),len(groups),nl,nc),
X_reconstructions.reshape(len(times),len(groups),nl,nc),
groups,
print_step=0)
max_abs_error: 6.7890625 mean_abs_error: 0.014291764742296214
/home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: divide by zero encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until /home/viluiz/anaconda3/envs/py3ml/lib/python3.7/site-packages/ipykernel_launcher.py:3: RuntimeWarning: invalid value encountered in true_divide This is separate from the ipykernel package so we can avoid doing imports until
In [110]:
X_train_encoded = conv_ae.layers[0].predict(X_train)
fig, ax = plt.subplots(1,1, figsize=[20,10])
ax.plot(times, X_train_encoded);
ax.grid()
ax.legend(range(15))
Out[110]:
<matplotlib.legend.Legend at 0x7ff80dff4650>
In [113]:
from tensorflow.keras.models import load_model
conv_ae.save("conv_ae.h5")
import joblib
joblib.dump(stdscaler, "stdscaler.pkl")
np.savetxt('X_train_encoded.csv', X_train_encoded, delimiter=',')
np.save('X_train.npy', X_train)
#...
# conv_ae = load_model("conv_ae.h5")
# stdscaler = joblib.load("stdscaler.pkl")
# X_train_compressed = np.loadtxt('X_train_encoded.csv', delimiter=',')
# X_train_1D = np.loadtxt('X_train_1D.csv', delimiter=',')
# times = np.loadtxt('times.csv', delimiter=',')
# with open('groups.txt') as f:
# groups = [g.strip() for g in f.readlines()]
# X_recovered = conv_ae.layers[1].predict(X_train_compressed)
# X_recovered = stdscaler.inverse_transform(np.moveaxis(X_recovered,3,1).reshape(len(times),len(group)*nl*nc))
In [71]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
im = ax.flatten()[i].imshow(X_reconstructions.reshape(len(times),len(groups),nl,nc)[100,i,:,:])
fig.colorbar(im, ax=ax.flatten()[i])
ax.flatten()[i].set_title(group)
In [72]:
fig, ax = plt.subplots(2,4, figsize=[20,10])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_reconstructions[:,i*nl*nc+4])
ax.flatten()[i].set_title(group)
In [93]:
fig, ax = plt.subplots(4,2, figsize=[20,40])
for i, group in enumerate(groups):
ax.flatten()[i].plot(times, X_train_1D[:,i*nl*nc+4])
ax.flatten()[i].plot(times, X_reconstructions[:,i*nl*nc+4],'--')
ax.flatten()[i].set_title(group)
In [ ]: